#dev 2020-04-19

2020-04-19 UTC
geoffo, gRegorLove, dckc and [snarfed] joined the channel
#
[snarfed] fluffy: bridgy support for private webmentions would be awesome! whether autoauth or whatever. i wouldn't lead it, since i don't really use private posts or wms myself, but i'd happily help anyone who wants to work on it
gRegorLove joined the channel
#
[snarfed] https://github.com/snarfed/bridgy/issues/11#issuecomment-490980323
#
Loqi [snarfed] cc @aaronpk @sknebel. we've been talking about reviving this idea and actually handling non-public responses (eg twitter replies from protected accounts) with https://indieweb.org/private_posts . see https://chat.indieweb.org/dev/2019-05-09#t15574194...
nickodd, Hypernought, MasonDiscord[m], bltavaresDiscord, JorropoDiscord[m, kitt, kanej[m], cesarosumDiscord, Exca1iburTheWise, gkimbwalaDiscord, xylanDiscord[m], berDiscord[m], rittmeDiscord[m], carsonfarmer[m], drshamoon[m], chrisDiscord[m], DamirDiscord[m], Nebulous[m], gabrielbaron16Di, MichaelTenDiscor, anthony-albertor, KinnardDiscord[7, SmileRobotDiscor, DoppelgngerDisco, robinzzzDiscord[, shivankDiscord[m, nyarlathotepDisc, vexlDiscord[m], AblibuDiscord[m], matyas_mustohaDi, pcowgillDiscord[, matschafferDisco, SomeguyDiscord[m, ReneM[m], AXEL-Brian[m], TristanDiscord[m, drshamoonDiscord, SpicoliWhiteDisc, BossMANDiscord[m, arjanvaneerselDi, dowlandaielloDis, HielleMatrixBrid, M123897974564Dis, romaric[m], llllllDiscord[m], corylDiscord[m], cesarosum[m], sekiDiscord[m], Sean|FortmaticDi, Canti0001[m], Luna14Discord[m], thesage1014Disco, chmanieDiscord[4, pranayDiscord[m], KeegenDiscord[m], SteelixDiscord[m, HyunwooLeeDiscor, kppDiscord[m], crestDiscord[m], ptonerDiscord[m], MikeShultzDiscor, MisterGoreDiscor, Senshi[m], EdEdorEddyDiscor, carsonfarmerDisc, HeysteinDiscord[, LSJI07Discord[m], sekiDiscord[m]1, peatDiscord[m], nrtxrmndDiscord[, OxyDiscord[m], vasa[m], SpaceOutlawDisco, gozala[m], catmanDiscord[m], dostDiscord[m], jmank88Discord[m, cristobalDiscord, JayWelsh0845[m], celsoDiscord[m]1, malaclypsDiscord, ShadowLingDiscor, RomaricDiscord[m, M3baidDiscord[m], HenniDiscord[m], foxcoolDiscord[4, drbh[m], SirMemesALotDisc, FranklinDiscord[, jmac, gRegorLove, vika_nezrimaya, [LewisCowles] and [jgmac1106] joined the channel
#
sknebel !tell aaronpk: re your validation question: on one hand I think it'd be nice to have pages with broken microformats show up as if they didn't have any, on the other hand I'm not sure if you can reliably tell them from pages with good mf2 and the link only appearing e.g. in a comment - would hinge on that for me
#
Loqi Ok, I'll tell them that when I see them next
#
sknebel TIL github has "template repositories" - https://help.github.com/en/github/creating-cloning-and-archiving-repositories/creating-a-template-repository - seems useful for quickstart website examples etc
#
[LewisCowles] egats
[mapkyca] joined the channel
#
jamietanna[m] Do we have anything that allows the previewing of Microformats? I'm adding MF2 support to someone else's site, and want to preview the newly created `h-event`.
#
jamietanna[m] I've used https://monocle.p3k.io/preview?url=https%3A%2F%2Fdeploy-preview-9713--devopsdays-web.netlify.app%2Fevents%2F2020-london%2Fwelcome%2F but it doesn't show too much
#
jamietanna[m] I've managed to get it loading in Indigenous for Android which gives me a nicer preview, but wondered if we had any alternatives
#
Zegnat Hmm, now that I know of. It is kind of up to every reader to see how it wants to represent mf2 data. There is no true 1 rendering
[jgmac1106] joined the channel
#
jamietanna[m] That's fair - I guess it's for some examples to show it off, preferably without it being in a reader. but np!
#
jamietanna[m] swentel has anything changed recently re image uploads in Indigenous? I've raised https://github.com/swentel/indigenous-android/issues/357 as I'm seeing two requests to my media endpoint - one to (successfully) upload the image, but then one to create a post, even when I'm using the `Upload` screen which I wouldn't expect a post to be created
#
Loqi [jamietanna] #357 Image upload "fails" to separate media endpoint
petermolnar joined the channel
#
Loqi aaronpk: sknebel left you a message 2 hours, 49 minutes ago: re your validation question: on one hand I think it'd be nice to have pages with broken microformats show up as if they didn't have any, on the other hand I'm not sure if you can reliably tell them from pages with good mf2 and the link only appearing e.g. in a comment - would hinge on that for me
#
aaronpk i should add events support to monocle tho
#
aaronpk sknebel: not sure i understand what you mean there
#
Loqi it is probable
#
sknebel it'd be not so good if the fallback would mean that e.g. a link in a webmention comment displayed on page would cause it to verify
#
aaronpk ahh
#
sknebel so accepting links outside the mf2 "main object" has some risk
#
aaronpk so, he wrote back: https://github.com/aaronpk/webmention.io/issues/150#issuecomment-616089717
#
Loqi [szpak] Thanks that you looked into that. The new error message is slightly more meaningful, but the reader still can be confused how to fix it. One of the articles found in the search engine is https://indieweb.org/microformats . It explains some things ...
#
sknebel maybe ignoring mf1 would be a compromise? I expect that to be way more common in broken in-the-wild examples
#
sknebel or some heuristic to recognize "broken" vs "properly formatted, but link is not in content", but that sounds difficult. but I don't have good insight into what actually happens
#
aaronpk one problem is the parser doesn't indicate whether data was found via mf1 or mf2
#
sknebel right, and you can't just turn it off
#
aaronpk so i'd have to check manually somehow, i guess a few xpath queries could do it
#
sknebel I don't have a good sense for how the noise of plain mentions would be. especially since I assume senders aren't commonly going to send mentions for links they don't intend to
#
sknebel maybe accept them and put some info in the output that site owners can decide to not render mentions where the url was outside the content if they find htem to be "bad quality"?
#
aaronpk i'm kind of liking the option of "if the link isn't found in the parsed mf content, and if there was only mf1, downgrade to a regular non-mf mention"
#
sknebel true, I guess that could be done with just searching for a h-entry class
#
aaronpk yeah i guess i have to check for lack of mf2 rather than presence of mf1?
#
aaronpk now i wish the parser would tell me whether it found mf1 or mf2
#
aaronpk ah the php parser can disable mf1!
geoffo, nickodd, petermolnar, [LewisCowles], [mapkyca], [jgmac1106], swentel, [jeremycherfas], andysylvester, deathrow1 and gxt joined the channel; nickodd left the channel