#KartikPrabhurepeating "saw your post-type-discovery algorithm draft https://www.w3.org/TR/2016/WD-post-type-discovery-20161028/#algorithm . Any particular reason for the ordering of types as "rsvp > reply > repost > like" ? I understand "reply" getting precedence but don't understand the "rsvp > reply" and the ordering after "reply""
#Loqitantek: KartikPrabhu left you a message 3 days, 18 hours ago: saw your post-type-discovery algorithm draft https://www.w3.org/TR/2016/WD-post-type-discovery-20161028/#algorithm . Any particular reason for the ordering of types as "rsvp > reply > repost > like" ? I understand "reply" getting precedence but don't understand the "rsvp > reply" and the ordering after "reply"
#gRegorLoveThe nested li was one thing, but I guess what I was thinking more was: there's a lot of redundant information with the full city/region listed under the shorthand
#ChrisAldrichPivotal is taking care of the lion's share. They're a relatively large company of programmers, so they've got lots of snacks/drinks in their kitchen area and will be catering lunch for us.
#ChrisAldrichThey've got a large open area in the front for the opening/closing/demos (with pingpong tables), and half a dozen or so individual highly connected conference rooms for breaking into groups of 3-15+ for sessions.
#sknebelhm, rereading the webmention spec the security section focusses on the sender side in a few subsections (don't send WMs to localhost, don't send WMs to servers in private networks), but the same concerns apply similarly to webmention verification?
#sknebel(although it's GET requests there, not POSTs)
#sknebelme too, and we've talked about it quite a bit
#sknebelso I was surprised that it isn't mentioned in the spec
#aaronpki think the concern is less important with GET requests, since normally GET requests don't mutate state. tho for really poorly designed systems i guess that could still be a problem.
tantek joined the channel
#sknebelon localhost, you potentially could talk to e.g. a database using a non-HTTP protocol, there have been weird vulnerabilities where creative HTTP requests then are interpreted as commands. especially relevant if you accept HTTP urls including port numbers
#sknebeland if verification results are displayed, you could leak info from internal sites?