#dev 2025-07-03
2025-07-03 UTC
# 
osteophage Saw a post about using Zola just earlier today.
# osteophage https://alavi.me/blog/why-i-used-zola/
# osteophage [edit] https://alavi.me/blog/why-i-used-zola/
grufwub and sebbu joined the channel
# Kolev Nice.
troojg, sebbu and barnaby joined the channel
# trwnh it might be great if that actually worked consistently... in most cases the http signature is from a proxy actor and not the user-associated actor > well it does enable fetching protected posts since it authenticates the requester
# trwnh technically you can choose to implement the outbox and/or inbox separately. the problem is that doing this isn't enough for fedi; you need to also implement a bunch of extra stuff like webfinger and http sigs cavage draft, and that's leaving out the issue of assumed schemas entirely... > activitypub is kind of all or nothing [...] it doesn't lend itself well to modular design
# trwnh in theory for just a reader, you can implement GET outbox (if the actor publishes an outbox and allows GET); this is almost directly analogous to an rss/atom feed except the data format is activities instead of entries, and you also probably have to deal with paging which differs across implementations > I don't want my reader to also have to implement ActivityPub pieces.
# trwnh the problem with this is that fedi expects specifically acct: URIs which require a localpart and a domainpart, since it represents a user account on a given service. "fun" fact, the localpart of an acct: URI can be an email address, not specifically a username :-) > it really should be simply `@domain`
# trwnh if i understand correctly the target use case of libravatar is to operates on email addresses rather than on websites (like gravatar before it), so it's not a 1:1 comparison. the problem is that everything in the space still uses gravatar directly and doesn't care about libravatar> libravatar is dead compared to rel=icon which is simpler and requires less power (HTML)
# trwnh this is a mastodon-specific url pattern, it doesn't need to be /@localpart and different softwares do different things (/localpart, /~localpart, /, it's completely arbitrary) > @example.com@example.com -> https://example.com/@example.com
# trwnh i know for a fact this will fail with pleroma/akkoma users
# trwnh snac/snac2 also doesn't follow this convention
# trwnh also you can host in a subdirectory, there is no requirement to serve on the root domain
# trwnh as far as i'm aware right now, only mastodon / misskey / gotosocial / mitra follow that convention
# trwnh this goes back to what you were saying about monoculture -- it only seems like a common pattern because most of the current fedi is mastodon/misskey/gts
# trwnh there are more softwares that do not follow this assumption than there are softwares that do. the "mastodon network" meme lives on in part due to assumptions like this. if people stopped using webfinger en masse and started doing a basic string manipulation instead, it would cut off a lot of softwares
# trwnh RFC 8820 section 2.3: "To avoid collisions, rigidity, and erroneous client assumptions, Specifications MUST NOT define a fixed prefix for their URI paths [...] unless allowed by the scheme definition"
# trwnh as much as you or i might not like or care for webfinger, a big part of the reason it's used is to smooth over all of these various arbitrary paths
# trwnh we could say "just make people remember their profile URI" but that didn't seem to work out well for openid sadly
# trwnh oh re: the gravatar thing, disqus uses gravatar too
# trwnh quasseldroid too
# trwnh twitter doesn't do /@user, it does /user. if you want to detect which software is running on which domain and maintain a mapping of @@ to urls, then okay... but you can't expect everyone to do that. at the end of the day the @@ or even the @ is just a personal convenience and not an actual universal identifier
# [snarfed] capjamesg re https://chat.indieweb.org/dev/2025-07-02#t1751497146578000 , glad you got it working! ...but that's definitely the hard way 😁 you were closer with get_activities_response. the standard way to get a feed, timeline, etc with granary is the get_activities method on any class
Saige_Leah and Maxpm joined the channel
# 
doesnm @capjamesg oh, are you rly want to make activitypub reader? i'm search this nowadays and don't found anything applicable except https://github.com/hyphacoop/reader.distributed.press
# 
Loqi A reader in the context of the indieweb is either a feature integrated into an indieweb site or a separate site or piece of software that provides a way to read content from indieweb sites and sometimes respond (like, comment, repost, etc) https://indieweb.org/reader
sarna joined the channel
[pfefferle] and jak2k joined the channel
jak2k, NaomiAmethyst, JadedBlueEyes, [KevinMarks], ttybitnik, jak2k1 and gRegor joined the channel
# 
gRegor Just gardened this section, I had no idea: https://indieweb.org/Gravatar#Web3_support
# 
[mattl] I forget now, but something like 30% of http://Libre.fm avatars were already on libravatar
# [mattl] http://Libre.fm is all no-javascript BTS fans
# [social] [mattl]++ for no-JS
barnaby joined the channel
[aciccarello] joined the channel
# 
[aciccarello] I link the photo on my hcard to my gravatar image. It makes it easier to update my profile in multiple places.
KE0VVT joined the channel
# 
capjamesg Is there a way to go from an arbitrary URL to figuring out if it is an AP profile (i.e. https://mastodon.online/@eeeps@front-end.social )?
# [aciccarello] I eventually want to have my static site builder cache external images which would fix the potential tracking issue.
ttybitnik, immibis, barnaby, NaomiAmethyst6, KE0VVT, angelo, gRegor and thegreekgeek__ joined the channel