mblaneyI could handle the encoding, but just wondering if it's intentional as I've never hit that issue during endpoint discovery before. (And I'm not sending any accept-encoding headers)
eli_oat, EmreSokullu, cweiske and tantek joined the channel
sknebelso HWC Berlin didn't end up in the newsletter because someone put "open end" as the end time and (surprise) the parsing script didn't manage to make a useful timestamp out of that. I'll fix the newsletter script to handle that more gracefully, but any suggestions how to publish something like that? "open end" as text, and have let'S say midnight in a machine-only mf2 attribute? /cc aaronpk
aaronpkI suspect it tried to convert that to a timestamp, which probably parsed to 1970, which then excluded it from the newsletter cause it’s in the past
sknebelyes, that's the issue pretty much. not having a end-property is covered, but not having one that can't be parsed to a timestamp. will fix for the newsletter, question remains if there is something we can do markup wise to make consuming apps do something better than "ignore property due to weird content"
ben_thatmustbemeits a good processing model to follow though, it future proofs it. Specs often say specifically you must ignore anything you don't understand, that way any future additions don't cause breakage
snarfed!tell mblaney looks like you have to send something like Accept-Encoding: identity to disable it...but that doesn't work here either. (i don't control this serving path; google does.) i'll see what i can do.
aaronpkGWG: you're on the once-a-day polling interval in case the foursquare web hook fails, have you waited at least that long to see if it catches up?
[eddie]aaronpk: Is there an easy way to get a token from indieauth.com? Compass wants a token, so my thought was to just build a quick app that let's you login via IndieAuth and it will present a token, but if there is an easier way then I'll do that
aaronpk"easy way to get a token from indieauth.com" isn't the right question. the token comes from the token endpoint which is entirely up to your website to decide how to do.
[eddie]Ahh gotcha. Thanks ? I guess I might sit down and see if I can put together an IndieAuth client into compass since I've done that for a couple projects now and if that works I can send a PR. If I run into issues I can use gimme a token as a backup if I'm not able to get it done before I head out somewhere
[eddie]aaronpk: Quick brainstorm to make sure I'm thinking down the right path. We would want to determine the micropub endpoint from the identity URL (same as any micropub app), but we also don't want to require you to post to the same identity you logged in with. (Example: Someone might log in with a static site identity but might want to post to their Known micropub endpoint). Which means you're starting the IndieAuth part from the very beginning. W
[eddie]URL? A "Micropub Endpoint" is obviously the Micropub endpoint itself, but we would want them to point to the homepage of whatever micropub server we want them to point to. From there, we would have them enter that URL, click "Login" and we would do the standard micropub app workflow to authenticate, get a token and discover the micropub_endpoint for that identity
[eddie]What do we call that URL? A "Micropub Endpoint" is obviously the Micropub endpoint itself, but we would want them to point to the homepage of whatever micropub server we want them to point to. From there, we would have them enter that URL, click "Login" and we would do the standard micropub app workflow to authenticate, get a token and discover the micropub_endpoint for that identity
[eddie]Okay, that makes sense, because we have the "Realtime Micropub Export" header, that provides the context for the "web sign-in". That sounds good. I was initially concerned web sign-in wouldn't make it obvious that they were doing their micropub, but the context on the page provides that for us.
tantekdidn't see bear post this so I will: Looks like WPA2 wifi is fundamentally (at the spec level!) crackable and any compliant implemenations are thus vuln https://www.krackattacks.com e.g. your home wifi access point
snarfedalso sounds like routers are less vulnerable than client devices, fortunately. "Our main attack is against the 4-way handshake, and does not exploit access points, but instead targets clients."
snarfedGWG: we do want new tests too though. the existing test timestamps without timezones were valuable. we want to test with them *and* with timezones
[miklb]I just meant that if it was a matter of using the client/enjoying your trip vs struggling with tests, use a fork with with a hotfix and address tests later. but sounds like you have a plan.
GWGsnarfed, if I keep using Micropub tools this week, my intention is to get many things fixed. I had originally thought that would be in my plugins only.
sknebel so HWC Berlin didn't end up in the newsletter because someone put "open end" as the end time and (surprise) the parsing script didn't manage to make a useful timestamp out of that. I'll fix the newsletter script to handle that more gracefully, but any suggestions how to publish something like that? "open end" as text, and have let'S say midnight in a machine-only mf2 attribute? /cc aaronpk
ben_thatmustbeme its a good processing model to follow though, it future proofs it. Specs often say specifically you must ignore anything you don't understand, that way any future additions don't cause breakage
[eddie] Ahh gotcha. Thanks ? I guess I might sit down and see if I can put together an IndieAuth client into compass since I've done that for a couple projects now and if that works I can send a PR. If I run into issues I can use gimme a token as a backup if I'm not able to get it done before I head out somewhere
tantek didn't see bear post this so I will: Looks like WPA2 wifi is fundamentally (at the spec level!) crackable and any compliant implemenations are thus vuln https://www.krackattacks.com e.g. your home wifi access point
aaronparecki.com edited /User:Aaronparecki.com (+1156) "/* working on */ indieweb community todo" (view diff)