#dev 2018-11-09

2018-11-09 UTC
#
@_mumblings
Replying to one of these usernames will send a Webmention to that user’s external web site. – Manton Reece, Usernames on http://Micro.blog That’s just fucking cool. https://www.manton.org/2018/11/08/usernames-on-microblog.html
(twitter.com/_/status/1060684864591220736)
snarfed, [tantek], eli_oat, [manton], sidedoor, renem, Vinmax, gRegorLove, cweiske, [sebsel], swentel and [jgmac1106] joined the channel
#
GWG
mblaney, still maintaining SimplePie?
#
mblaney
hi GWG yes but I've made so many changes to it recently I'm not sure I should push to main.
#
GWG
mblaney: Just wondering if you think I should just bundle my own version when I need it, since I can't get the WordPress one updated.
#
mblaney
that's probably a bit ironic but not sure existing user base want my h-feed changes.
#
GWG
mblaney: Actually, people want an update who don't care about h-feed.
#
GWG
Where is the changelog anyway?
#
mblaney
oh yes I think you should bundle your own. mainline is much better than what wordpress has.
#
GWG
That's all I needed to know, other than where to find the changelog. I can't find it
#
GWG
I want to know what has happened since 1.3.1
#
GWG
mblaney: My Hack Day project at IWC Berlin was using SimplePie to create an RSS to JF2 converter to support a Microsub endpoint that is being developed.
#
GWG
mblaney: I will read the commits either way. Any pointers you might have re my goal?
#
GWG
The current code works. Just want to improve its quality
#
GWG
My h-feed parsing code is separate, so not sure I need that support.
#
mblaney
that's quite the project GWG.
#
GWG
mblaney: I had working code by the end of the day
#
mblaney
what is your goal GWG?
#
Loqi
[dshanske] parse-this: Parse This Parsing Library for WordPress- Can Act as a Standalone Plugin
#
GWG
mblaney: It is for a Microsub endpoint
#
GWG
Goal is turn any URL into structured jf2 data.
#
GWG
Similar to aaronpk's x-ray
#
mblaney
are you working on yarns?
#
mblaney
I'm not sure what pointers I can give you, just pulling in and using your own SimplePie will fix a few parsing issues.
#
mblaney
but if you want a full microsub server, then SimplePie can also do caching for you, so I would use it for microformats too.
#
mblaney
what is yarns?
#
Loqi
Yarns Indie Reader or Yarns for short, is an reader that is built for use with WordPress as a plugin https://indieweb.org/Yarns
#
GWG
mblaney: Not exactly. Yarns is using my parsing code.
#
GWG
I'll have a look at the SimplePie h-feed code.
#
GWG
mblaney: Can I pass a DOMDocument directly into SimplePie?
#
GWG
I already have one. But I just looked quickly, and you can't turn off the MF2 parsing. And converting to SimplePie then back to microformats seems like extra work when I already have it.
#
GWG
But I'll figure it out.
#
mblaney
you can't pass DOMDocument but you can pass html that will be loaded into it.
#
GWG
I know. Trying to optimize.
#
GWG
I'll see what I am thinking after I read and try to use the code. I could always open an issue/file a PR if I want something.
#
mblaney
sure happy to look at it.
#
mblaney
sorry that will take some digging through, but if you do want more microformats updates in mainline SimplePie let me know.
#
GWG
mblaney: I'm always happy to contribute.
#
GWG
If I find something that I think people other than me might want
#
GWG
Who is using the mf feature of SimplePie? Is there a lot of interest?
deathrow1 joined the channel
#
mblaney
just me I think, SimplePie is pretty much a parked project otherwise.
#
GWG
Just bugfixes?
#
GWG
So, I might think of some simple things
#
GWG
But I doubt anything earthshattering
eli_oat and [grantcodes] joined the channel
#
eli_oat
ahoy! I am working on writing a new micropub endpoint. A feature I'd like to add is the ability to know what micropub client sent a request to the endpoint, is there a standard way of identifying them?
#
[grantcodes]
what is x-app?
#
Loqi
It looks like we don't have a page for "x-app" yet. Would you like to create it? (Or just say "x-app is ____", a sentence describing the term)
#
[grantcodes]
what is h-app?
#
Loqi
h-x-app is a proposed microformats vocabulary for marking up data about software applications https://indieweb.org/h-app
#
[grantcodes]
I think you need to use something like that eli_oat
#
eli_oat
ah, and do any micropub clients support this?
#
eli_oat
ah, I see that quill does
#
eli_oat
groovy
#
eli_oat
thanks!
#
[grantcodes]
Yep, need to update all my stuff to use it at some point
#
[grantcodes]
I think Zegnat and others were doing research into apps also using a manifest.json file, but not sure where that was documented
#
[grantcodes]
what is manifest.json?
#
Loqi
It looks like we don't have a page for "manifest.json" yet. Would you like to create it? (Or just say "manifest.json is ____", a sentence describing the term)
#
[grantcodes]
what is web manifest?
#
Loqi
It looks like we don't have a page for "web manifest" yet. Would you like to create it? (Or just say "web manifest is ____", a sentence describing the term)
[tantek] joined the channel
#
sknebel
what is Web App Manifest?
#
Loqi
Web App Manifest is a W3C living document that defines a JSON sidefile for web applications to publish information such as their name, icon, description https://indieweb.org/Web_App_Manifest
#
eli_oat
ya'll are amazing! thank you
#
[grantcodes]
I was so close!
eli_oat joined the channel
#
sknebel
chat search helps in such cases I find
#
aaronpk
eli_oat: before you can use h-app or whatever, there's a step before
#
aaronpk
the Micropub request of course includes an access token, and you have to have previously issued that access token. That process will have included the client_id in the request
#
sknebel
manifest.json is either a [[Web App Manifest]] or a key file in a [[WebExtension]].
#
sknebel
what is manifest.json?
#
Loqi
It looks like we don't have a page for "manifest.json" yet. Would you like to create it? (Or just say "manifest.json is ____", a sentence describing the term)
#
sknebel
manifest.json is either a [[Web App Manifest]] or a key file in a [[WebExtension]].
#
aaronpk
so at the very least you can always show the client_id that made the request, and since it's a URL it will be useful on its own
#
sknebel
what is client_id?
#
Loqi
It looks like we don't have a page for "client_id" yet. Would you like to create it? (Or just say "client_id is ____", a sentence describing the term)
[tmiller], [eddie], eli_oat, snarfed and [Vincent] joined the channel
#
[Vincent]
also manifest.json can be defined as manifest.webmanifest https://developer.mozilla.org/en-US/docs/Web/Manifest
#
[Vincent]
but like no one uses it lol
#
Loqi
[Vincent]: [eddie] left you a message 3 days, 4 hours ago: thanks, I’ll take a look!
eli_oat joined the channel
#
aaronpk
I just got followed by 6 mastodon spam bots
#
sknebel
from normal instances, or spam-instances?
#
[Vincent]
why do bots follow you? I get why they post things obviously. Is it just to appear human?
#
aaronpk
from a single instance
#
sknebel
to make you look at their profile
#
[Vincent]
ah
#
[Vincent]
I see
#
aaronpk
same as instagram follow spam I guess
#
[Vincent]
If social sites charged a small entry (one off fee). I suppose it would significantly reduce the bots overnight. Even 1 Euro would probably do it.
[Rose] joined the channel
#
[Rose]
It would however probably force teenagers who parents don't want to pay onto free instances mostly filled by spam bots.
#
[Rose]
(And a chunk of adults who think everything on the internet should be free)
#
aaronpk
the internet already charges a small fee (domain name registration)
#
aaronpk
whoever set up this spam instance is paying for the domain
#
[Vincent]
true. I suppose the point I was making, was that if it wasn’t economically viable. They wouldn’t spin up tons of bots. A domain name is for multiples I assume, so at least the cost scales to a point
#
[Vincent]
but then cost creates its own Silo as Rose pointed out
#
aaronpk
right, but my point is that even though there is already a cost to run spam, it is often worth it to the spammers. increasing that cost to the point that it's no longer worth it will also very likely increase it beyond a reasonable amount for a legit person
#
[Vincent]
I suppose the person who solves that wins the internet
#
aaronpk
haha yeah
#
aaronpk
guess I need to figure out how I want to implement blocking
#
sknebel
it probably shifts the type of spam - e.g. those Bitcoin scams on Twitter go to the lengths of hacking verified accounts, or buying hacked verified accounts, but if one succeeds it makes quite a bit of money
#
sknebel
wasn't there some forum site that charged $5 per account?
#
aaronpk
I remember one that asked you to mail in $1 in an envelope
#
sknebel
I think there was one that just wanted a postcard
#
aaronpk
minimal blocklist successful
#
sknebel
oh, mediawiki moves keep the pageid intact... that's kind of surprisin
[tantek] joined the channel
#
aaronpk
alright time to figure out this missing author info issue
#
Loqi
[aaronpk] #79 Missing author info on some mf2 feeds
[jgmac1106] joined the channel
#
[jgmac1106]
[aaronpk] see if the missing author issue is always on posts with an h-cite... Maybe same issue that happen here is happening elsewhere...often see missing author on like and bookmarks
#
aaronpk
nah it's just interesting combinations of where the author info appears
#
aaronpk
i'm documenting it on that issue
[manton], eli_oat, tin and [kevinmarks] joined the channel
#
aaronpk
damn, not sure my logic is correct after all
#
aaronpk
if there is a single h-entry and a single h-card on a page at the same level, and the h-card has a URL that is not that page, but the h-entry does not have a URL, is that page a feed or a permalink?
#
aaronpk
now i'm leaning towards the answer being a permalink
#
aaronpk
that's what adactio.com does on permalinks
#
sknebel
would being able to handle rel=bookmark correctly help?
#
sknebel
probably has edge cases too
#
sknebel
there's a reason explicit h-feed exists...
#
aaronpk
hm i'm not sure how rel=bookmark fits into this
#
aaronpk
also ugh adactio's server doesn't respond to Accept headers right
#
aaronpk
I think it's only looking at the first one
#
sknebel
rel bookmark is actually scoped to elements on the page
#
aaronpk
oh god, tell that to the microformats parsers
#
sknebel
even without that, rel=bookmark != page url is an indication that it is not a permalink
#
sknebel
but I guess people could put shorturls or other alternative urls there...
#
GWG
aaronpk, looking at your problem...that explains why I have so much trouble
#
aaronpk
this is the challenge of letting people be so flexible with markup
#
aaronpk
on one hand, it's great that we don't have to be so prescriptive, and people can structure stuff how they want. on the other hand, gawd it makes a mess of trying to understand all that.
#
sknebel
yeah. I think pushing people to wrap stuff in an h-feed is fair
#
aaronpk
I think I have this the way I want now
#
GWG
I would appreciate it if they wrapped things in a h-feed
#
GWG
How do I ask nicely?
#
aaronpk
it will only "break" on the case where someone is intending to provide a feed but the feed contains only one item
#
aaronpk
(which incidentally, is how some of my tag pages work!)
j12t joined the channel
#
GWG
Why do some people use h-feed?
#
aaronpk
oh no haha
#
aaronpk
after all this, adactio's h-card on his feed pages doesn't include a photo
#
aaronpk
gRegorLove: how attached to your home page markup are you? you have an h-card that appears to be an item in your h-feed, rather than the author of the h-feed. if you just add "u-author" to the list of classes on your h-card it works great.
#
aaronpk
that's the only time i've seen that pattern, and I think I would rather ask you to fix it instead of coding around it :)
#
gRegorLove
Not attached. :) I think that's a bug; it's supposed to be outside the h-feed, similar to your homepage
#
gRegorLove
Hm, maybe I never finished that. I thought I'd switched to `u-author=/` on my posts
#
aaronpk
hmm outside the h-feed? I don't think I've seen that one either
#
aaronpk
my home page doesn't have an h-feed, it's just a list of h-entrys followed by an h-card
#
gRegorLove
yeah, that's what I mean, just a top-level h-card on my homepage.
#
aaronpk
with a sibling h-feed??
#
gRegorLove
The h-feed is a separate thing. I wanted to get rid of p-author h-card on each individual post, instead using u-author=/, with a footer h-card on the homepage.
#
aaronpk
sure getting rid of that makes sense
#
aaronpk
but why do you want the main h-card outside of your h-feed?
#
gRegorLove
Maybe it can stay there? I need to refresh on /authorship
#
aaronpk
I mean right now literally the only change you need to make is to add the u-author class
#
gRegorLove
The h-feed wasn't a consideration at the time I was working on that.
#
aaronpk
seems like it'd be more work to move things around
[eddie] joined the channel
#
gRegorLove
Not disagreeing. Just need to make sure it still works with /authorship from post permalinks, which I think it will.
#
gRegorLove
Because it has u-url u-uid
snarfed, sebsel, j12t and [eddie] joined the channel
#
@jgmac1106
Feel like watching the #el30 talk between @downes and @benwerd is a "http://Brid.gy Too Far" in The Caine-Hackman Theory in our #ds106 and #IndieWeb comparative case study. (http://quickthoughts.jgregorymcverry.com/s/1PP7Pc)
(twitter.com/_/status/1060996757293854720)
j12t, [jgmac1106] and [manton] joined the channel
#
[manton]
Mentioning this as I work on ActivityPub-related stuff in case other people have run into it: Mastodon's RSS feed item "permalinks" do not match the format used from ActivityPub requests, e.g. /@username/123 vs. /users/username/statuses/123. I'm having to translate these URLs in some cases to avoid duplicates, which seems wrong.
#
aaronpk
that doesn't sound right, I am pretty sure mastodon doesn't care about what the permalinks look like
#
aaronpk
I didnt have to do any fiddling like that for my site
#
aaronpk
oh wait, I misread, you are consuming the mastodon RSS feeds?
#
aaronpk
lol mastodon's RSS and Atom feeds use different URLs even
#
aaronpk
their RSS uses the /@username/123 format, but their Atom uses /users/username/statuses/123
#
[manton]
Oh wow.
#
[manton]
Actually, in that case I can switch to Atom and solve part of this.
#
aaronpk
why do you need to parse their rss/atom at all?
#
Zegnat
Do they 301? That is: they represent the same resource?
#
[manton]
Since everything else in M.b is based on feeds, when you first look up an ActivityPub user, I pull in posts from their feed if there is one. Not sure I will keep it that way but it kind of automatically worked and seemed nice.
#
aaronpk
don't they have an activitypub outbox for that?
#
[manton]
The /@username/123 URLs do redirect. I'm really not sure why there are 2 versions.
#
aaronpk
the links don't redirect for me
#
[manton]
Oh wait, they don't redirect.
#
[manton]
Maybe. They redirect _to_ the @username version, looks like.
#
aaronpk
the pages also have microformats so that's another option https://pin13.net/mf2/?url=https%3A%2F%2Fmastodon.social%2F%40Gargron
#
Loqi
@DearMsDear
#
[manton]
[aaronpk] Thanks. I don't pull from the outbox at all currently. I guess I could.
[eddie] joined the channel
#
[manton]
So the MF2 version matches the RSS permalinks. Atom and ActivityPub use the other format. Feels like they changed their mind and just didn't update everything.
doubleloop joined the channel
#
doubleloop
Testing the authorship algorithm on my homepage, but it seems to fail at step 2
#
doubleloop
I checked the microformats in micrometa.jkphl.is and there's an h-entry with an author property with an embedded h-card
#
doubleloop
(inside an h-feed)
#
sknebel
it doesn't handle feeds
#
sknebel
so limitation of the tool
#
Zegnat
Also limitation of the algo, see question raised re feeds under https://indieweb.org/authorship#Questions
#
sknebel
not sure why it even shows step 2, that's also wrong
#
sknebel
well, not quite. it could look at the entries in the feed, and run for those
#
sknebel
just never added that
#
doubleloop
Thanks sknebel, Zegnat
#
doubleloop
aaronpk: I think my h-feed also has the issue as gRegorLove's. In that the h-card is an item of the h-feed, rather than an author property.
#
doubleloop
I think that's coming from the fact that SemPress has h-feed on the body tag, and I have a little widget with my h-card in my sidebar.
#
doubleloop
I added u-author to the h-card in the sidebar now, that looks better
#
doubleloop
Oops, now I have two h-cards for the author on individual post pages though. Will that cause a problem?
#
aaronpk
The test tool should show you the result on the post permalinks, I think it will be fine
#
aaronpk
You can also test a post page in xray to see what it thinks
#
doubleloop
Looks OK, it picks up the first one. I have two slightly different h-cards... one being generated by the semantic linkbacks plugin for WP (I think) per entry, and one from my sidebar widget.
#
doubleloop
But that's another issue. Thanks!
j12t joined the channel
#
gRegorLove
I noticed that bc of some Loqi previews of Greg's links
#
aaronpk
I don’t even understand what that page is supposed to be
#
aaronpk
the url on that h-card is not the url of the page so the h-card is discounted from being the representative item on the page
j12t joined the channel
#
aaronpk
I guess XRay should parse it as a feed of h-cites but I don’t really have a consuming use case for that so it doesn’t make sense