#dev 2023-02-24

2023-02-24 UTC
Skyther[d], IWDiscordRelay and petermolnar joined the channel
#
[tantek]
sure, to Mastodon, that's my point. This is a quirk of Mastodon's inbound federation support
#
[tantek]
regardless of whether it comes from BF or directly from a personal site like aaronpk's
#
[tantek]
seems like the current theory (e.g. from aaronpk's example) is that a Mastodon requires an AS2 hashtag entry to at least have: "type": "Hashtag",
#
[tantek]
unknown if it also needs an href (though possibly, in order to replace the tag link in the content with an instance-local tag link)
gRegor joined the channel
#
gRegor
Yeah, and the spec issues I linked yesterday
#
gRegor
May not be AP spec like I said there, though. In the AP #235 discussion cwebber said it would be an independent extension (of AS2, I think?).
#
gRegor
Regardless, it does not appear to be standardized yet, but what Mastodon implemented.
#
gRegor
These are the type of things I was going to summarize
geoffo, gxt_, IWSlackGateway, AramZS, gRegor, [jeremycherfas] and gerben joined the channel
#
@luctielen
↩️ I want to look into webmentions for mine, but idk how hard that is going to be.
(twitter.com/_/status/1629017992368357378)
[aciccarello] joined the channel
#
[aciccarello]
[capjamesg] I'm looking at updating the docs on webmention.app (at least in my fork). I saw you had opened this issue and was wondering if could clarify what you initially tried/expected. https://github.com/remy/wm/issues/43
#
Loqi
[preview] [capjamesg] #43 Add new guidance to "How to integrate with Netlify" section of the documentation
gRegor, gxt__, AramZS, Guest6_, robotcat, IWSlackGateway, CrowderSoup, geoffo and [dave] joined the channel
#
Loqi
[preview] [caseywatts] #186 Ignore Query Parameters?
[snarfed] joined the channel
#
[snarfed]
definitely can't just drop them all since they're meaningful on some sites
#
[snarfed]
I guess you could special case tracking params like utm_*
#
aaronpk
yeah it's a weird one
#
[snarfed]
or I think you could reasonably decline. wm target values aren't really the best place to do that kind of URL cleanup
#
[snarfed]
(or source values)
#
sknebel
well you could offer a switch for the api to ignore all for sites that know they can be ignored
#
sknebel
although it feels like they should(TM) be cleaned up earlier....
#
sknebel
but I guess thats difficult for wm.io which can't know
[aciccarello] joined the channel
#
[aciccarello]
Is there any way currently to query those mentions if you don't know the exact query params?
#
aaronpk
right now there's nothing special about query parameters, a URL is a string and may or may not have query params
#
[aciccarello]
Maybe webmention sending should check for the canonical url?
#
sknebel
sending needs to send the url as it appears on the page
#
sknebel
verification could, *if* it can know the canoncial url (easy if the endpoint is integrated with the site or the site redirects, harder otherwise), store the mapping to the canonical url
geoffo and [tantek] joined the channel
#
[tantek]
ignoring query params would/could be a different API
#
[tantek]
e.g. targetprefix=
#
[tantek]
that way the caller could decide if they care about extra stuff on the end of the URL or not
#
[tantek]
better to keep 'target' pretty strict though as in that precise URL
#
[tantek]
target= that is, as the API param to wm io
[schmarty] joined the channel
#
aaronpk
hmm actually i think i've gotten that prefix request before too
geoffo and gRegor joined the channel
#
gRegor
yikes
geoffo joined the channel
#
[aciccarello]
I've wondered how much damage extensions could do
#
[aciccarello]
Might be time to review who maintains the ones I use 🤔
#
aaronpk
this is why if you publish an extension and it has any level of popularity, you immediately start getting emails asking if you'll sell it
#
[snarfed]
mobile apps too
#
aaronpk
less so on iOS but yes
[KevinMarks], [benatwork], [pfefferle] and robotcat joined the channel