KevinMarks, KartikPrabhu and [pfefferle] joined the channel
#petermolnarso... I revisited this json-ld thing, because of all the hype activitypub is getting... apparently json-ld data has to be inlined. This is one of the most idiotic engineering decision I've ever seen: besides duplicating the data, like content, you HAVE to duplicate it within the same document.
#sknebelactivitypub doesn't inline, it uses separate feeds
#sknebelGoogle wants JSON-LD inlined for their SEO stuff, cause they don
#jgmac1106_hopefully the weather holds and the kids go to the pool so I can record later
#jgmac1106_I will say seeing how Snopes used the custom form fields in WordPress for Claim/Review was pretty cool. Also note only paid partners get to know the authors of reviews. This was noted as both a monetization and security strategy.
#jgmac1106_I just meant how the fields were used and populated not so much the json-ld approach to credibility…..so reminded me of the early days of credit card payments and everybody wanting a diffeerent logo and no consumer caring
#jgmac1106_but I spend many hours trying to teach young kids to evaluate claims and evidence so seeing their workflow in WordPress and how form fields could be connected to any kind of metadata at UI level was neat
#snarfedhi oodani! figured it might be easier to talk granary here than on twitter
#oodaniah, so json_to_object works when mapped over the individual h-entries, and then there's a make_activities_base_response call at the end?
#snarfedyes, but you probaby don't need make_activities_base_response
#snarfedthat's for serving an API response over HTTP
#oodaniah, fair enough. i imagine django-rest-framework will handle that part for me :p
#oodanidoes granary work with paginated feeds? i noticed the activitystreams output on granary.io has "items per page" fields, but it doesn't seem to have a field for the page's rel="next" attribute
#oodanii'm not sure how activitystreams handles pagination so. maybe it works and i just don't understand it yet?
#snarfedyou include startIndex and/or count query params in the HTTP request
[miklb] joined the channel
#snarfedah, no, i see what you mean. if you give granary an HTML page as input with rel=next or rel=prev, it doesn't currently know to follow those. that'd be a great addition! PRs welcome!
#oodanii do know i want granary to be outputting atom - on the live version of my site i'm using django's feed framework and it produces /really weird/ xml? <link> tags with a </link> instead of a />, stuff like that
#oodaniand it looks like to get atom out of granary, i need to put activitystreams into it :p
#oodanisending my html to granary.io produces pretty nice atom already, so i could technically just render to html and then have granary parse that, but it'll be quicker if i can keep everything in dictionaries :p
#snarfedalso maybe consider just using django's, if it's ok otherwise! separate </link> closing tags in atom may not be as pretty as self-closing, but they are valid, and harmless
#snarfedalso, re latency, i expect the network round trip to/from granary swamps the time for your site to serve html vs another format
#oodanithe atom from granary is also "enriched" with activitystreams types and such, which django's isn't, and i think it'd be useful to have that extra info?
#[miklb]if someone distributes a self-signed cert for local dev on https, and someone were to put that cert on a site with malicious code, and I’ve authorized it locally in my browser, then navigate to malicious site, my browser would accept the cert, yes?
#oodanithere's a few other prerequisites for that kind of attack - in particular, the attacker would need to poison your dns in order to have the malicious site use the same domain name as is on the certificate - but yep
#cweiskeif you lose your own CA certificate then you have a problem
#cweiskebut not if you only lose the domain certificate
#Loqi[jgarber]: sknebel left you a message 5 days ago: curious why a bunch of packages for different endpoints instead of one generic rel=x-finding one?
#Loqi[jgarber]: sknebel left you a message 5 days ago: curious why a bunch of packages for different endpoints instead of one generic rel=x-finding one?
#Loqi[jgarber]: sknebel left you a message 5 days ago: curious why a bunch of packages for different endpoints instead of one generic rel=x-finding one?
#Loqi[jgarber]: sknebel left you a message 5 days ago: curious why a bunch of packages for different endpoints instead of one generic rel=x-finding one?
#Loqi[jgarber]: sknebel left you a message 5 days ago: curious why a bunch of packages for different endpoints instead of one generic rel=x-finding one?
#[jgarber]Good question, sknebel. I believe I wrote the individual gems in such a way that they could easily be rolled up by a top-level meta gem that makes a single HTTP request and parses the response using whichever combination of `rel` values you might be looking for.
#[jgarber]…or, you could include just the one gem for the one `rel` value your project cares about.
#[jgarber]The use case is arguable, to be certain. But it makes for some small, easily testable gems.
#[jgarber]Tangential to that, what is interesting (to me, at least…) is the difference in Webmention’s and Micropub’s endpoint discovery algorithms:
#Loqisnarfed has 70 karma in this channel (406 overall)
#sknebelwhich is why I asked the question: generally, you can use the same rel-finding code for all our use cases unless I'm totally forgetting about something.
#[cleverdevil]!tell [manton] I'm getting pretty close with a functional version of a Facebook -> Micro.blog toolkit. The only hangup is that Micro.blog doesn't seem to be respecting the `published` property of the MF2 JSON that I am sending across via Micropub, so all of the content ends up with the wrong date/time of publish.
#[manton][cleverdevil] Micro.blog definitely supports `published` with form-encoded params because we use that for the Instagram import, but yeah, must be missing from MF2 JSON. Should be an easy fix.
#sknebelof course it has, just like Trackbacks. you can put spammy links on other people's sites and then a few months later beg them to delete it to get rid of your google penalties for spammy backlinks
[jgmac1106], KartikPrabhu, snarfed and benwerd joined the channel
#jgmac1106_[gRegorLove] back to the question of is Css Grid an equal serving of div soup than bootstrap. Almost equal for me now but there is a bunch of shorthand I will be able to use in the future when I understand that I think will lead to cleaner HTML
tantek__, [pfefferle], snarfed and [miklb] joined the channel
#LoqiInternationalization (AKA internationalisation, i18n; localization, l10n.) is the process of adapting software/content to various languages https://indieweb.org/i18n