#dev 2018-04-05

2018-04-05 UTC
chrisaldrich, snarfed, KartikPrabhu, [cleverdevil], j12t, petermolnar_, renem, jmac_, deathrow1|absnt, petermolnar, [jgmac1106], tantek, aaronpk and [aaronpk] joined the channel
#
Loqi
/credentialing should be deleted for being abstract and not relevant to indieweb per reasons in the page.
#
Loqi
Countdown set by tantek on 2018-04-03 at 11:29am PDT
#
tantek
ty loqi
snarfed joined the channel
#
tantek
algorithmic feed << 2018-04-02 Benedict Evans: [https://www.ben-evans.com/benedictevans/2018/4/2/the-death-of-the-newsfeed The death of the newsfeed]
#
Loqi
ok, I added "2018-04-02 Benedict Evans: [https://www.ben-evans.com/benedictevans/2018/4/2/the-death-of-the-newsfeed The death of the newsfeed]" to the "See Also" section of /algorithmic_feed
#
tantek.com
deleted /credentialing "content was abstract concept with no current relevance to the IndieWeb, no active IndieWeb examples or brainstorming towards active implementation"
#
tantek.com
edited /Indieweb_for_Education (-20) "credentialing--"
(view diff)
#
tantek.com
edited /badge (-20) "credentialing--"
(view diff)
#
tantek
there do seem to be many references to credentials though
#
tantek
on the wiki
#
tantek
purely about login / auth credentials though
#
tantek
which makes sense, and is directly applicable to much indieweb
#
loqi.me
edited /Slack (+257) "tantek added "2018-04-02 John O'Nolan: [https://threadreaderapp.com/thread/980872508395188224.html After 4 years of running a 11,000 member public Slack community ... we’ve decided to shut the whole thing down for good.] - thread unrolled. Criticism:..."
(view diff)
renem joined the channel
#
loqi.me
edited /Slack (+317) "tantek added "https://twitter.com/JohnONolan/status/980873578764697600" to "See Also""
(view diff)
#
www.boffosocko.com
edited /embargo (+1659) "example; see also"
(view diff)
#
loqi.me
edited /Facebook (+407) "tantek added "Criticism: growth at all costs: https://twitter.com/cwarzel/status/979478092077973504" to "See Also""
(view diff)
#
loqi.me
edited /Facebook (+440) "tantek added "2018-03-30 Criticism: Blocks fbexposed URL: https://twitter.com/MelissaRyan/status/979769434867994625" to "See Also""
(view diff)
#
loqi.me
edited /social_media (+159) "tantek added "2018-03-30 NYT: [https://www.nytimes.com/2018/03/30/world/americas/travelers-visa-social-media.html 14 Million Visitors to U.S. Face Social Media Screening]" to "See Also""
(view diff)
#
tantek.com
edited /vk (+27) "link to name, fix dfn, linky, capitalizations"
(view diff)
#
tantek.com
moved /vk to /VK "canonical caps"
[cleverdevil] joined the channel
#
www.boffosocko.com
edited /use_case (+337) "Links to business use cases"
(view diff)
bdesham and snarfed joined the channel
#
tantek.com
edited /Google (+61) "/* Dead Silos */ link Google Wave, and support article on it"
(view diff)
[miklb] joined the channel
#
cleverdevil.io
edited /Internet_Archive (+99) "/* IndieWeb Examples */"
(view diff)
#
loqi.me
edited /meetup.com (+59) "aaronpk added "https://aaronparecki.com/2018/04/04/15/" to "See Also""
(view diff)
#
www.boffosocko.com
edited /IndieWeb_for_business (+1386) "William Hertling; Brainstorming, Reviews, acquisition posts; See Also section"
(view diff)
snarfed joined the channel
#
Loqi
ok, I added "https://github.com/Wiredcraft/jekyllplus" to the "See Also" section of /Jekyll
#
Loqi
[Wiredcraft] jekyllplus: Lightweight CMS for GitHub pages and Jekyll websites.
chrisaldrich and [miklb] joined the channel
#
tantek.com
edited /Facebook (+228) "/* How to export your data */ link to cleverdevil's blog post"
(view diff)
snarfed and [ryan339] joined the channel
#
loqi.me
created /Microsub_bridge (+147) "prompted by tantek and dfn added by tantek"
(view diff)
#
kaja.sknebel.net
edited /Microsub_bridge (+1) "linkify ('x is y. <url>.' pattern)"
(view diff)
#
tantek
Microsub << Microsub bridge
#
Loqi
ok, I added "[[Microsub bridge]]" to the "See Also" section of /Microsub
snarfed joined the channel
#
tantek
j12t proposed a coop model for a Google Reader replacement
[snarfed] joined the channel
#
tantek
we're done here in SF!
#
Loqi
[aaronpk] #24 don't generate the authorization code until after the user clicks approve
[miklb] joined the channel
#
aaronpk
That's gonna be a tricky one
#
GWG
If I have it redirect back to the same URL that shows the authorize page, with an extra parameter that bypasses the screen and instead, just redirects to the code, is that an issue as someone could try to force that?
#
GWG
Right now, the initial endpoint redirects to a second URL, which requires someone be logged into a site.
#
aaronpk
Do you have to use a get request for that? Can you make it only generate the code on a post request instead?
#
GWG
Should I be verifying that the second URL was redirected from the first by checking the server redirect?
#
GWG
I'll have to check the code.
#
GWG
Yes, I can.
#
GWG
aaronpk: I'm really knocking them out tonight.
j12t and j12t_ joined the channel
#
GWG
aaronpk: How do I solve #29?
j12t and tantek joined the channel
#
GWG
Never mind. Think I thought of something.
#
GWG
No, that didn't pan out. Question still remains.
#
GWG
The only thing I can think of is changing the token endpoint to not return a token if the scope isn't set.
#
GWG
But nothing on the authorization side.
barpthewire, tantek, KartikPrabhu, AngeloGladding, [kevinmarks], [colinwalker] and loicm joined the channel
#
[kevinmarks]
This sounds like an indieweb friendly job http://www.fogcreek.com/jobs/GlitchDesignEngineer
Loqi, Kongaloosh, raucao, aaronpk, schmarty, peterlawson[m], grantcodes, [pfefferle] and loicm joined the channel
#
microblog.ducamp.me
edited /Facebook-fr (+49760) "synchro with original (translation + hyperlinks to be reviewed/completed)"
(view diff)
#
microblog.ducamp.me
edited /Facebook-fr (+11) "/* Comment se sevrer soi-même */ typo"
(view diff)
[colinwalker], [kevinmarks], [pfefferle] and [jgmac1106] joined the channel
#
@Lobbicyclette
@perlkonig Hey, I just discovered your Grav Webmention plugin. Do you know of a theme that uses it ? I'd love to see how to implement it on my website.
(twitter.com/_/status/981866831739805696)
#
skippy
i may be in the minority here, but I find the loqi content in here very disruptive to the conversation. :( i dont want to completely ignore loqi, because then i might miss out on !tell messages left for me. :(
#
skippy
but all the wiki edits are less-than-useful.
#
Zegnat
The idea is that people can go and review the edits, which is especially important for brainstorm pages and on-wiki discussions.
#
Zegnat
I guess this is one of those “how do you use indieweb.org” issues again.
#
skippy
sure.
#
aaronpk
We could talk about moving all wiki edits to meta. Right now just the event edits are in meta
#
skippy
i'm the new guy here, so i dont want to try to overly impose my own preferences. I just find it hard to follow a thread, or read a back scroll, while trying to manually visually filter out all the loqi content spread between the humans chatting.
#
scripter.co
edited /User:Scripter.co (+134) "/* Site details */"
(view diff)
#
skippy
i dont idle in meta, so moving it all over there would be fine with me; until sucj time as I have a meta conversation to have and need to chat there ;)
#
Zegnat
You mean a meta chat like the discussion about what wiki edits should show in which channel? ;)
#
skippy
*zing*
kaushalmodi joined the channel
#
kaushalmodi
I agree with skippy. Ironically I just made few edits. Would be nice if they ended up in meta.
#
Zegnat
I wonder if we should add a note to the wiki interface that checking “minor” makes an edit not show in chat.
#
Zegnat
(sknebel should Kaja do its autolinking etc as a minor edit?)
#
skippy
kaushalmodi: why did you select hugo for your site generator? curious to learn what other people's motivations are for the tools they select.
#
sknebel
Zegnat: could do, since I think it can be trusted now - early on it was intentionally not, to catch mistakes
#
sknebel
(if your client has good filtering features, you probably could exclude the wikiedits only if you find them really annoying)
#
skippy
yes, I could learn how to filter just the wiki edits via irssi. i can do that, if the consesus is that most people find real value in the edit announcements here.
#
skippy
like i said, I dont want to force my preferences. just expressing them right now, to see if i'm the outlier.
#
Zegnat
No, I think it is a good discussion to bring up skippy! Especially if we start to more strictly refer conversations from #indieweb to here, it is important not to break up the flow.
#
Zegnat
That is one of the reasons we limited Loqi output in #indieweb in the first place :)
leg joined the channel
#
aaronpk
speaking of which, this conversation should take place in -meta :)
[miklb] joined the channel
#
skippy
when POSSEing notes to twitter, do folks generally capture the Twitter URL after it's been posted, for inclusion on their original copy?
#
aaronpk
then display it on the post with a "u-syndication" class
#
skippy
ah! okay.
#
aaronpk
I have little icons that appear next to the timestamp of my post with all the syndication links
#
skippy
that leads to another question I've had about all the indieweb markup, generally.
#
skippy
all this markup is to make things machine-readable and parsable for various uses. I'm curious what the use case is for machine-consuming all the syndication links, etc...
#
aaronpk
helps with deduping posts when they are received at the other end
#
aaronpk
if you look at some replies on my posts, they will also have syndication link icons
#
aaronpk
so if I get a webmention from an indie URL and it says it has a twitter URL syndication, then when I get the bridgy webmention for that I can match it up and avoid showing the reply twice
#
loqi.me
edited /test (+25) "aaronpk added "hello" to "See Also""
(view diff)
singpolyma joined the channel
#
skippy
kaushalmodi: why did you select hugo for your site generator? curious to learn what other people's motivations are for the tools they select.
#
GWG
aaronpk, I understand that the token endpoint shouldn't accept codes without scope, which would prevent response type id from issuing a token. But how does the authorization endpoint do anything?
#
aaronpk
what do you mean?
#
aaronpk
the authorization endpoint will need to associate the requested response type with the code it issues
#
aaronpk
that way the token endpoint can check what response type was expected from the code and reject it unless it's "code"
#
GWG
aaronpk, how does the token endpoint check that?
#
aaronpk
you know how you're storing a bunch of stuff along with the code already?
#
aaronpk
like the expiration date
#
aaronpk
just add response_type to that array
#
GWG
aaronpk, it is added
#
aaronpk
okay then when the token endpoint is checking to see if the code is valid, it can look at that property as well
#
GWG
aaronpk, right now the two act independently
#
GWG
The token endpoint doesn't directly verify, it verifies as if they weren't integrated
#
aaronpk
oh? oh I see.. it actually makes a POST request to itself
#
GWG
Yes.
#
aaronpk
"If the authorization code was issued with no scope, the token endpoint MUST NOT issue an access token, as empty scopes are invalid per Section 3.3 of OAuth 2.0 [RFC6749]."
#
GWG
aaronpk, oh... I did that last night.
#
GWG
I thought there was something the authorization endpoint had to do
#
aaronpk
so you're fine then as long as the token endpoint rejects an authorization code with no scope
#
GWG
Okay. Does that mean I am ready to request a merge?
#
aaronpk
want me to look at it again?
#
GWG
Please
#
GWG
One last time.
KartikPrabhu joined the channel
#
aaronpk
for future reference, it's helpful if you include a link to the issue you're solving in the commit message
#
GWG
Good point
KartikPrabhu, tantek and kaushalmodi joined the channel
#
kaushalmodi
skippy: About why I chose Hugo.. I had got burnt by trying to get Octropress installed, and then stay away from Ruby stuff ever since. Then about a year back, I came across Hugo on one of the HN posts, and the "one static binary" really appealed me. Additionally installing the Go toolchain was simply extracting an archive and putting that in your PATH. With Hugo, I simply upload the static built binary of its Dev version and build my site usin
#
KartikPrabhu
what is Hugo?
#
Loqi
Hugo is a static site generator written in Go https://indieweb.org/Hugo
#
kaushalmodi
.. build my site using that. The benefits of awesome speed, community support, awesomeness of the Go templating which resembles sort of to Lisp got realised later, and I am glad I picked up this tool.
[kevinmarks] joined the channel
#
[kevinmarks]
The speed is a big thing, but you usually only see that if you're importing an already big site.
#
skippy
importing all my tweets into my site has raised my build time from < 1 second to ~4 seconds on my macbook; and ~20 seconds on my server.
#
tantek
is each tweet in a separate file?
#
skippy
$ find . -type f | wc -l 12243
#
skippy
paste fail. 12243 files total.
#
tantek
wow ok that may explain that. individual file overhead tends to be an I/O (thus perf) bottleneck
#
skippy
20 seconds to generate 15K pages is pretty darned fast, so I can't complain too much. :) Either spend resources generating each request on demand, or spend resources generating the whole site with each occasional addition of new content. :shrug:
[kevinmarks] joined the channel
#
petermolnar
20 seconds for 15k pages is close to insane
#
petermolnar
and impressive
#
kevinmarks.com
created /tf;dr (+215) "prompted by petermolnar"
(view diff)
snarfed joined the channel
#
aaronparecki.com
edited /site-deaths (+328) "/* Upcoming */ foodspotting"
(view diff)
kaushalmodi joined the channel
#
kaushalmodi
skippy: As I said on the Discourse forum, you can make that time insane(r) if you publish 10 or so tweets per page :)
#
skippy
yes; but that introduces complications for micropub, and including them in time-order along with my posts.
#
kaushalmodi
without knowing anything about micropub, would it help if each of those pages were an h-feed
#
kaushalmodi
and each of the tweets an h-entry?
#
skippy
i'm still wrestling with how i want to display my content...
#
kaushalmodi
and the u-url of each h-entry would be the link to the same page, but with fragment portion.. like (foo.com/#tweet1)
#
skippy
there . pushed up my latest build to my site for public view https://skippy.net/
#
Loqi
Scott Merrill
tantek joined the channel
#
skippy
havent decided if i actuall WANT to display all my tweets that way or not.
#
snarfed
skippy: many of us keep tweets (etc) on our sites but de-emphasize them by just not showing them on the front page
#
tantek
it's not either or, there's a third option which is lazy regeneration, regenerating pages only upon first request, and then caching that static file
#
snarfed
^ yes! i do that
#
snarfed
skippy: e.g. https://snarfed.org/ generally has my original tweets, but not replies, RTs, etc. those are on https://snarfed.org/responses
#
skippy
Hugo is like Jekyll:build the whole site.
#
kaushalmodi
skippy: You can prebuild the tweet pages and move them to static
#
skippy
ah yes. i could.
#
aaronpk
!tell adactio your webmention endpoint seems to have disappeared!
#
Loqi
Ok, I'll tell them that when I see them next
#
kaushalmodi
If you are planning to keep using twitter, that flow might become cumbersome
#
skippy
that makes it harder to list the tweets on the home page, though.
#
kaushalmodi
But if this is a one time thing, it would work great
#
skippy
i am not planning to ditch Twitter. But I would like to post to my site, and syndicate to twitter. IDeally with micropub.
#
sknebel
aaronpk: probably better contacting him somewhere, haven't seen him in IRC for quite a while
#
kaushalmodi
I haven't tried, but using the File methods should work: https://gohugo.io/templates/files/
#
skippy
kaushalmodi: yes, i just read that bit, too.
#
KartikPrabhu
aaronpk: I see it in his omepage head
#
aaronpk
oops jkphl's webmention endpoint is sending back a redirect to https
#
aaronpk
where did we land on what a sender should do in that case?
#
aaronpk
the endpoint is advertised as http://example.com/wm but sending a POST request there returns an HTTP 301 redirect to https://example.com/wm
#
aaronpk
IIRC youre not supposed to repeat a POST request after receiving a 301
#
tantek
301 in practice doesn't support POST
#
tantek
303 does AFAIK
#
tantek
worth a test in the test suite
#
tantek
since clearly it's easy enough to make that error
#
tantek
had a thought re: SWAT0 and IWS, how about trying to get micro.blog's indieweb support to the point where it can place at least one of the A,B,C roles by IWS?
#
aaronpk
ooh I wonder what that would take
#
tantek
right, it's good to have high level goals like that - it's also demoable
#
tantek
and it challenges other software and/or providers to do the same
#
aaronpk
I think it might be able to be C right now https://indieweb.org/SWAT0#C_must_support
#
tantek
e.g. how long does it take to setup a new WordPress instance so that it can do at least one of the roles A,B,C?
#
tantek
or Known for that matter?
#
tantek
we got SWAT0 working in 2015 with Gen 1 level tools. Can we get SWAT0 working with Gen 2 level tools now?
#
aaronpk
I don't think many things support person tagging at all yet
#
tantek.com
edited /SWAT0 (+21) "jargon"
(view diff)
#
tantek
petermolnar: re: https://news.ycombinator.com/item?id=16764602 in particular this is one way fail to ever produce a spec: "... I suspect we're better off thinking about all possible features, because this allows us to define a protocol that generalizes better."
#
Loqi
[amelius] Emphasis should be on the requirements. What do we want the system to do? And how can we control our privacy? Let's answer these questions first.
#
tantek
"thinking about all possible features" = you will never ship
#
tantek
"what's a minimum feature set for a social network"? Twitter 2006
#
tantek
aaronpk I wonder what happened with that - person tagging
#
tantek
what is person tagging
#
Loqi
A person tag (AKA people tag) is a person mention that is also a tag on a post that refers to a specific person by URL rather than just a word or phrase, and is done as an explicit tagging action by the user, beyond just mentioning a person via hyperlink / h-card / or @-name, autocompleted or not https://indieweb.org/person-tagging
[jgmac1106] joined the channel
#
tantek
2 min timebox expired on yet another handwavy Hacker News discussion of decentralization.
#
snarfed.org
edited /syndication-models (+316) "brainstorming: whether and when to automate"
(view diff)
#
GWG
Re person tagging, is that a challenge?
#
snarfed
it is now :P
#
GWG
Syndication Links has it
j12t and tantek joined the channel
#
www.boffosocko.com
edited /multi-site_indieweb (+736) "Patrick Rhone reasoning"
(view diff)
[snarfed], chrisaldrich, snarfed and barpthewire joined the channel
#
@spigot
↩️ This is the output and format from http://Brid.gy. Not so good. The link is there but what the heck is the rest? I’m working on getting webmentions working on the site, this is a bit of a setback.
(twitter.com/_/status/981957032537305088)
#
KartikPrabhu
[snarfed]: ^ ?
#
aaronpk
looks like bad microformats on the page being published
#
KartikPrabhu
what does that have to do with bridgy?
#
aaronpk
it looks like bridgy's fault from the user's perspective
#
aaronpk
(bridgy publish)
#
KartikPrabhu
<sigh> it is wordpress. I'm not even going to touch that bad microformats issue on that
#
aaronpk
of course it is
#
sknebel
yeah, he's added mf2 classes but the mf1 classes are still there and backcompat triggers and ruins it
#
Zegnat
turns off backcompat parsing
#
Zegnat
Aaah, thats better
#
aaronpk
weird, backcompat parsing isn't supposed to override mf2 classes
#
sknebel
different elements
#
KartikPrabhu
aaronpk: it does in new mf2py
#
KartikPrabhu
which isn't released :(
#
sknebel
it's wordpress mf1 elements nested into manually added mf2
#
KartikPrabhu
slowly backs away from wordpress
#
sknebel
<body h-feed> <div hfeed> <content> *boom*
#
sknebel
eh, <body h-entry>
#
Zegnat
feed in entry. I guess there might be a valid case for that? Maybe?
#
KartikPrabhu
Zegnat: can have a comments feed I suppose
#
KartikPrabhu
anywho wordpress microformats is not something I want to address
#
Zegnat
It could have, but then those comments can’t be marked as u-comment on the h-entry anymore
#
KartikPrabhu
Zegnat: p-comment h-feed with a list of h-entry
#
Zegnat
Interesting idea
#
KartikPrabhu
list of h-cite too
#
GWG
aaronpk: You did all the Errors in the Authorization Endpoint. Should I do the ones in the Token Endpoint as well with the same format?
#
aaronpk
oh looks like i missed a couple ther
#
GWG
I'll handle it
#
aaronpk
yes please
#
GWG
What is the code for a redirect_uri not being on the same host or whitelisted?
#
Loqi
It looks like we don't have a page for "code for a redirect_uri not being on the same host or whitelisted" yet. Would you like to create it? (Or just say "code for a redirect_uri not being on the same host or whitelisted is ____", a sentence describing the term)
#
aaronpk
missing parameters are invalid_request, invalid stuff in the parameters are invalid_grant
[chrisaldrich] and j12t joined the channel
#
GWG
aaronpk: I'm asking for a review from pfefferle
[jjdelc] joined the channel
#
[jjdelc]
Hey guys, I've been working on this MP client for the last couple of months https://grumble.isgeek.net/ I've been using it with my (also homebrew) mp server
#
GWG
aaronpk: Next on to scope for Micropub
#
GWG
Limited scope
[miklb] and snarfed joined the channel
#
Zegnat
Interesting [jjdelc]! Can anyone use it at that domain?
#
Zegnat
Hmm, it seems to hang for me on “Initializing...”
#
aaronpk
GWG: so the micropub plugin needs to be aware of the scope of the token that's making the request
#
GWG
aaronpk: It is. It just doesn't do anything with it.
#
aaronpk
it looks like right now it's only aware of the user that the token belongs to
#
aaronpk
oh okay cool
#
aaronpk
i see there are already some checks for whether the user can create/update/delete posts
#
aaronpk
so you can just add some more checks that make sure the token has the needed scopes
#
GWG
I'm setting up a scope function that is passed the permission and determines if it is in scope.
#
GWG
It has to be called for each type of function
#
GWG
I've written it, but the unit testing is an issue
Kongaloosh joined the channel
#
[jjdelc]
Zegna https://github.com/jjdelc/grumble here's the source, but yeah the app is pure client side so it should work for any MP endpoint
#
Loqi
[jjdelc] grumble: Micropub client
#
[jjdelc]
Zegnat, I had some of that recently too, can you try refreshing? I think there's some problems sometimes loading VueJS
#
Zegnat
Trying in a new private window, see how it does without cache
#
skippy
not working for me at all with Firefox.
#
[jjdelc]
ah but it does on chrome... it looks like something with the service workers according to the console
#
skippy
button doesnt do anything; pressing "enter" doesn't work.
#
[jjdelc]
🤦‍♂️ will look at that
#
Zegnat
I am seeing the fetch for the URL I enter, and then nothing, in my Firefox
#
Zegnat
tries Safari
#
Zegnat
Oh hang on, I think both Firefox and Safari are hanging on cross-origin
#
Zegnat
What is CORS?
#
Loqi
CORS is an acronym for "cross-origin resource sharing," a mechanism for allowing browsers to make JavaScript requests to fetch resources from other domains https://indieweb.org/CORS
#
[jjdelc]
The client does need CORS to be enabled on the MP endpoint though
#
[jjdelc]
but if it's hanging on "Initializing" that's a different issue
#
[jjdelc]
I made a change, looks like I didn't have `vue.js` registered with UpUp for it to be available as offline resource
#
[jjdelc]
btw, the Micropub spec says nothing about CORS on the micropub endpoint, for my client to be full client side, I needed to enable CORS on my backend
#
Zegnat
I’d say the Micropub spec expects a client to have the ability to access resources over HTTP. JavaScript in browsers doesn’t do that. Not really a spec problem.
#
Zegnat
Though can you catch the CORS error? Maybe display an error to the user that they need CORS enabled on their endpoint for this to work?
#
[jjdelc]
I could do that, yeah
#
Zegnat
Hmm. And probably not only the endpoint either, but also on the page where you run endpoint discovery. As well as on the IndieAuth endpoints. Yikes.
#
[jjdelc]
I will not discover that until after successful login
#
[jjdelc]
yes, it all needs to be CORSed
#
[jjdelc]
since I do a client side read of the login page to discover the `authorization_endpoint`
#
[jjdelc]
Also the media endpoint
#
sknebel
[jjdelc]: you can add it to the list at https://indieweb.org/Micropub/Clients
#
[jjdelc]
sknebel will do, but I want to fix that firefox situation first 😕
#
skippy
this looks very interesting to me, [jjdelc]. thanks!
#
[jjdelc]
skippy did it load fine for you?
#
skippy
not yet
#
[jjdelc]
when I do a hard refresh from firefox it loads fine, but following a link it breaks
#
grantcodes
[jjdelc]: Cool! Nice to see more new micropub clients!
#
grantcodes
[[jjdelc]++
#
Loqi
jjdelc has 2 karma
#
[jjdelc]
Just pasting the url on a private window works fine https://grumble.isgeek.net/
#
grantcodes
I also got stuck "initializing" on firefox
#
skippy
Source map error: request failed with status 404
#
skippy
Source Map URL: upup.min.js.map[Learn More]
j12t joined the channel
#
Zegnat
Source maps shouldn’t matter, iirc
#
skippy
i know nothing about JS. just sharing what Firefox dev tool console is showing me.
[gerwitz] joined the channel
#
[gerwitz]
Does anyone have a favorite means for deciding a Micropub post is “photo” (e.g. from OwnYourGram) instead of “note that contains a photo” (e.g. from Micro.blog clients)?
#
grantcodes
Yeah source maps won't cause issues (still good to fix though)
#
grantcodes
The issue I'm having is with my own token endpoint returning an error for some reason
#
grantcodes
[jjdelc]: Might be worth having a look at my micropub helper library: https://github.com/grantcodes/micropub/ it's pretty solid these days
#
reliable.servesarcasm.com
created /User:Reliable.servesarcasm.com (+528) "Created page with "<h3>Hi!</h3> <p> I'm Dylan Harris. I've had a website up for a couple of decades. It's all my own cat food. Don't go visiting it, it's running on an Intel chip so old it's im...""
(view diff)
#
[jjdelc]
grantcodes, cool will look at that
#
[jjdelc]
Although maybe there is value in keeping parallel implementations
#
grantcodes
Absolutely, feel free to grab parts from it though :)
#
Zegnat
[gerwitz], yes, post type discovery
#
Zegnat
What is post type discovery?
#
Loqi
Post Type Discovery specifies an algorithm for determining the type of a post by what properties it has and potentially what value(s) they have, which helps avoid the need for explicit post types that are being abandoned by modern post creation UIs https://indieweb.org/post-type-discovery
#
grantcodes
The other issue you'll run into with a js client is you can't read headers, some people may have their endpoint links in headers.
#
[jjdelc]
oh
KartikPrabhu joined the channel
#
[jjdelc]
actually fetch() responses do include the headers
#
[jjdelc]
I realized that I'm using the response Location on the client
#
[jjdelc]
but yeah, at the moment I'm only disconvering the links from the HTML attributes
#
aaronpk
[gerwitz]: micro.blog posts are HTML with img tags, whereas native photo posts in micropub have a photo parameter
#
aaronpk
this popped up in my reader and i had a slight moment of panic and then realized it's correct :) https://media.aaronpk.com/file-20180405130418-6920.png
#
KartikPrabhu
ha! yeah just saw that on my reader too
#
skippy
aaronpk: you advised earlier to capture the twitter URL of POSSEd content, and link to it from the original with u-syndication. Curious about workflow: do you post the original content to twitter first, then capture that tweet url, then include that in your content that you post to your site? or do you post to your site, submit to twitter, then edit the original?
#
KartikPrabhu
I suspect the later. which is also what I do
#
aaronpk
mine is all automated, so I first post to my website with a micropub client, my site stores the post, then it POSSEs to whatever places were selected, and when it does that it writes the URL back to the post when it's done
#
skippy
gotcha. thanks. for a database-powered site, I see the utlity in that. for my current efforts to use a static site generator, I'm leaning toward the other model, to minimize local writes / regenerations.
#
aaronpk
mine isn't database-powered, it's just generated on the fly
#
gregorlove.com
edited /safety_status (+64) "/* gRegor Morrill */ another example"
(view diff)
#
KartikPrabhu
skippy: I do the same with files
[sdepolo] and [kevinmarks] joined the channel
[kaushal_modi] joined the channel
#
[kaushal_modi]
skippy: I do the same (publish the site twice): 1. Publish site 2. Post to twitter, get the tweet id 3. Add id to the post front-matter and republish
#
bear
skippy: I also do the same, I maintain all webmention, likes and other items as secondary files containing the info. If those secondary files are touched, I rebuild the articles associated with it
#
KartikPrabhu
also don't think file vs database is that different in this case, is it?
#
bear
I don't think so - it's more of: where do you store the metadata for things and can it be updated with changes propogated to your generated output
tantek and [gerwitz] joined the channel
#
[gerwitz]
@zegnat and [aaronpk]: thanks, I needed a refresher on some obviousness, there!
#
skippy
thanks everyone!
KartikPrabhu, snarfed, tantek, [colinwalker], j12t, klandwehr, leg and oodani joined the channel