#microformats 2017-05-01
2017-05-01 UTC
davidmead and [eddie] joined the channel
# 
ben_thatmustbeme mf1 wasn't case sensative at all was it?
davidmead and gRegorLove joined the channel
# 
gRegorLove I'm still confused on the <script> <style> e-parsing. I recall Proposal 2 being agreed to at IWC last year. http://microformats.org/wiki/microformats2-parsing-issues#exclude_style_elements_before_parsing
# gRegorLove The 'html' value keeps it, the 'value' doesn't
# gRegorLove "Proposal 2: Drop both <script> and <style> elements completely when parsing any property (except for e-* HTML values, which preserve all markup)."
tantek, [jeremycherfas], [colinwalker], [manton] and [kevinmarks] joined the channel
# 
@codepo8 CSS Grid in production - https://cssgrid.design/ - no big players yet, reminds me of microformats sites repos, this can be spammed soon. (twitter.com/_/status/859017912492994560)
# @codepo8 @kevinmarks The point in the Microformats example is that people added one rel=me in the footer to get spam sites i… https://twitter.com/i/web/status/859021726814015493 (twitter.com/_/status/859021726814015493)
# 
@kevinmarks @codepo8 If a spammer adopts grid (or microformats) should that be blocked from an examples site? (twitter.com/_/status/859020898736779264)
# 
benthatmustbeme edited /microformats2-parsing-faq (+10) "/* which backcompat root classes */ broken section links" (view diff)
# ben_thatmustbeme wat? includes?
# ben_thatmustbeme oh god
# ben_thatmustbeme schmarty: thats only if you want to deal with backward compatibility for v1
tantek, [cleverdevil], gRegorLove, JonathanNeal, KartikPrabhu, [eddie] and hober joined the channel
# ben_thatmustbeme '137 examples, 42 failures' and i think many of the failures are dt- related
# ben_thatmustbeme there are definitely other issues as well, like the backcompat list doesn't actually list hitem hfeed or hnews, but they are all used, so i don't know what properties to parse for
# ben_thatmustbeme also my PR keeps growing https://github.com/microformats/tests/pull/59/
# ben_thatmustbeme pretty sure this is wrong, it shouldn't treat adr as p-adr h-adr unless its inside a backcompat class
# gRegorLove Similar issue, yeah. Agreed, ben_thatmustbeme, the 'adr' should be ignored there.
# ben_thatmustbeme well, not completely, it should be a child h-adr
# ben_thatmustbeme as its still a root class
# ben_thatmustbeme err backcompat root class
tantek joined the channel
# gRegorLove Oh, right.
# gRegorLove I think upgrading it to p-adr h-adr is correct. http://microformats.org/wiki/h-card##adr+-+parse+as+p-adr+h-adr
# gRegorLove I think that preserves the intent of the mf1 adr
# gRegorLove Hm, scratch that. I'm not sure.
# gRegorLove 'adr' as property and root mf1 takes me a while to understand each time I think about it.
# gRegorLove It's always an mf1 root, sometimes also an mf1 property of h-card
# ben_thatmustbeme Microformats parsers SHOULD detect classic properties only if a classic root class name is found and parse them as microformats2 properties.
# ben_thatmustbeme yeah,
# ben_thatmustbeme in that instance, its inside a mf2 root, so it should be as a child h-adr, not p-adr h-adr
# ben_thatmustbeme hmm, damn, so if you are inside a back-compat root, you shouldn't recognize any modern p-*, dt-*, etc
# ben_thatmustbeme thats another thing that example brings up
# gRegorLove Correct.
# ben_thatmustbeme okay
# ben_thatmustbeme grumbles back to the drawing board
# gRegorLove It took me a while to get that right in my latest PR. I know you're not wanting to look at other parsers, but the backcompat() method might help.
# ben_thatmustbeme nah, i think i have it
# gRegorLove Some good test HTML for that in https://github.com/indieweb/php-mf2/issues/45 and https://github.com/indieweb/php-mf2/issues/107
# ben_thatmustbeme wow, yeah, comparing those to the output of some other parsers will certainly help
# ben_thatmustbeme don't have any properties for h-items
# ben_thatmustbeme because there was nothing specified in backcompat section of spec
# ben_thatmustbeme i think i found another bugin the mixed folder
# ben_thatmustbeme the h-card should have a value of "Aaron Pareki" since it has a name:
# ben_thatmustbeme wow, okay, yeah, definitely was comparing the output of the example you gave me gRegorLove to parse results from pin13 without actually really looking at it
# ben_thatmustbeme definitely not the thing to compare to, lol
# ben_thatmustbeme guessing its not on the new version of the code yet
# ben_thatmustbeme i think date parsing is the only part left thats got an error in my mixed folder
# ben_thatmustbeme that and the one issue i said with h-entry/mixedroots
# gRegorLove Yeah, no new php-mf2 release yet
# ben_thatmustbeme though i think unmung is wrong too, http://www.unmung.com/mf2?url=https%3A%2F%2Fw7apk.com%2Fkenwood-th-d74a&html=&pretty=on
# ben_thatmustbeme has both author and reviewer
# ben_thatmustbeme not sure
# 
KartikPrabhu yes
# KartikPrabhu this is backcompat h-review?
# gRegorLove ben_thatmustbeme: For php-mf2/45, this is the improved backcompat https://gregorlove.com/php-mf2/test.php?id=8
# gRegorLove And for php-mf2/107: https://gregorlove.com/php-mf2/test.php?id=2
# KartikPrabhu hmm why is unmung adding h-item to the reviewer?
# gRegorLove Whoa, hello Firefox JSON viewer
# gRegorLove I think I still prefer the raw view
# KartikPrabhu I Like the JSON viewer
# gRegorLove The indentation of the values is just weird to me
# gRegorLove KartikPrabhu: "reviewer" shouldn't be parsed in that ^^
# KartikPrabhu gRegorLove: why? it has a classic mf with reviewer property
# KartikPrabhu aaronpk: yeah the chrome one does look nicer with the "level indicators" (vertical lines)
# gRegorLove KartikPrabhu: It parses as p-author in backcompat
# KartikPrabhu oh! I see
# KartikPrabhu kylewm wrote a lot of that backcompat stuff in mf2py so I have little idea about it. Maybe file a bug?
# ben_thatmustbeme hmm, need a good way to compare some json easily
# ben_thatmustbeme rake spec just dumps out way too much of it
# ben_thatmustbeme hmm, some discrepancies in w7apk (107) parse gRegorLove
# ben_thatmustbeme reviewer - parse as p-author, including compat root vcard in the absence of h-card
# ben_thatmustbeme oh wait, hmm
# ben_thatmustbeme summary parse as p-name
# ben_thatmustbeme that one you don't seem to be doing
# ben_thatmustbeme this a huge task though huh, i have put in quite a lot of hours to this parser
# ben_thatmustbeme far more than i expected
# gRegorLove Yeah, it's a lot trickier than it looks at first glance
# gRegorLove What's the discrepancy on reviewer? I see author property as type h-card
# gRegorLove Weird regression on summary to p-name; could have sworn I had that one. Thanks!
[eddie] and [kevinmarks] joined the channel
# 
[kevinmarks] unmung is mf2py, but I did tweak it a little
# [kevinmarks] I think I added some bonus hfeed compatibility, as well as my gratuitous support for the BBC's recipe markup
# [kevinmarks] hm these comments are a bit worrying
# [kevinmarks] 'hreview': {
# [kevinmarks] 'summary': ['p-name'],
# [kevinmarks] # doesn't work properly, see spec
# [kevinmarks] 'fn': ['p-item', 'h-item', 'p-name'],
# [kevinmarks] # of the item being reviewed (p-item h-item u-photo)
# [kevinmarks] # of the item being reviewed (p-item h-item u-url)
# [kevinmarks] 'photo': ['u-photo'],
# [kevinmarks] 'url': ['u-url'],
# [kevinmarks] 'reviewer': ['p-reviewer', 'p-author', 'h-card'],
# gRegorLove A lot of that seems to correlate with php-mf2 backcompat parsing at the time
# gRegorLove It was basically a search-and-replace for old properties, not taking into account whether inside a backcompat root or mf2 root
# [kevinmarks] ah. mf2py is a bit like that but i think it handles nested beter
[mko] joined the channel
[eddie] and [kevinmarks] joined the channel
# [kevinmarks] Gregorlove++
[tamaracks] joined the channel