#microformats 2018-03-13
2018-03-13 UTC
tantek, [snarfed], [kevinmarks], [cleverdevil], [eddie], [asuh], nitot, [tantek], barpthewire, globbot, [Mike], [miklb], TallTed and Garbee joined the channel
# 
KartikPrabhu [kevinmarks]: sknebel: for your review https://github.com/microformats/mf2py/pull/98
tantek joined the channel
[eddie], KartikPrabhu, [kevinmarks], tantek, [miklb] and hober2 joined the channel
# 
gRegorLove fyi aaronpk, I'm working on https://github.com/indieweb/php-mf2/issues/149 now
# 
Zegnat Re #indieweb-chat and other date vagueness. If anyone wants to open an issue to discuss how they would expect date handling in the spec that would be swell. I think I discussed that before but do not have the time :( I did get some whitespace-linebreaks-stuff written up in travel downtime that I hope to put in for consideration tomorrow :)
# gRegorLove Any more dt- edge cases to throw at this while I'm making updates? It's no longer normalizing TZ offset when in a datetime attribute -- only in VCP.
# gRegorLove Very carefully :)
# gRegorLove I think with this fix we're pretty good on mf2 dt- parsing (easy) and even on VCP (much more confusing).
# gRegorLove Eh, they're working pretty well in the majority of cases, just hard for implementers to wrap their mind around. Something simpler might be nice, but I don't know if it's worth a rather big spec rewrite.
# gRegorLove Not opposed to the discussion or going that direction though, just my initial thoughts.
# KartikPrabhu VCP rules are mostly for backcompat no? why revise them entirely?
# KartikPrabhu also I think mf2py does vcp correctly now so it is not impossible
# KartikPrabhu I think the VCP spec language is pretty clear apart from some "which text to get issues"
webchat86 joined the channel
# KartikPrabhu I don't see how they are related. we can normalise on VCP since it is basically putting together a datetime. But in usual dt-* the datetime is actually authored
# KartikPrabhu that looks fine to me
# KartikPrabhu you could also put random words like "bundungle foo bar"
# KartikPrabhu who cares
# KartikPrabhu beause that is what VCP is intended for. It is tailored to use cases
# Zegnat I guess this is a better comparison. The exact same ABBR, but under VCP it has a completely different parsed outcome: http://php.microformats.io/?id=20180313224002737
# KartikPrabhu ha! that should not happen. mf2py only picks up the mf2 dt-* since the VCP does not match spec
# KartikPrabhu so that is another bug in phpmf2
# KartikPrabhu aah ok yes
# KartikPrabhu good
# KartikPrabhu I also like "publdished" as a property ;)
# KartikPrabhu according to vcp parsing rules
# Zegnat What? No. http://microformats.org/wiki/value-class-pattern#Basic_Parsing - step 2.3.
# KartikPrabhu hmm intersting
# KartikPrabhu yeah from another reading that looks right
# KartikPrabhu Zegnat: time for you to post issues to phpmf2 and mf2py :)
# KartikPrabhu and others
[cleverdevil] joined the channel
# KartikPrabhu sknebel: depends on the current owner
# KartikPrabhu and if someone is willing to take it
# gRegorLove I'm confused by those examples and what's expected.
# KartikPrabhu gRegorLove: apparently accodring to spec, if there is a single value element no vcp-dt parsing rules are to be used and the value is directly used
# KartikPrabhu Zegnat: maybe open an issue on the parsing rules repo and tag people
# gRegorLove Ohh, I was getting confused by both being in the same h-entry. This makes it clearer to me: https://php.microformats.io/?id=20180313230558077
# KartikPrabhu gRegorLove: yes
# gRegorLove Which yes, is corrected in master: http://pin13.net/mf2-dev/?id=20180313230707113
# gRegorLove Not an issue afaik
# gRegorLove VCP has strict formats for matching date parts
# KartikPrabhu it isn't technically corrected. Since if there is only one value then it should be used
# KartikPrabhu gRegorLove: see rule 3 http://microformats.org/wiki/value-class-pattern#Basic_Parsing
# KartikPrabhu "where there are multiple descendents..."
# KartikPrabhu mf2py gets this wrong too
# KartikPrabhu ha!
# KartikPrabhu night Zegnat
# gRegorLove Yeah, seems ambiguous when I read "For all date time properties (as defined in their respective microformats specifications), the following rules apply in addition to (and in some cases replacing) the above value class pattern parsing rules." under Date and Time Parsing.
# gRegorLove In which cases is it "replacing" vs just "in addition to". I think I've always read it as "replacing"
# KartikPrabhu yeah me too (even though I only read this part a few days ago! :P)
# gRegorLove http://microformats.org/wiki/value-class-pattern##if+the+element+is+an+abbr+element reads to me like it replaces basic parsing
sebsel joined the channel
# gRegorLove since it covers the cases (abbr, data, img) mentioned in basic parsing
# gRegorLove Let's just simplify it to everyone has to publish Swatch Internet Time :)
# Zegnat I agree it is ambiguous on the replacing part. The mf2 spec however says “parse the element for the value-class-pattern, including the date and time parsing rules”, which makes me think dt-* should adhere to the basic spec and not by default overwrite. And according to the basic vcp rules, the “date and time parsing rules” that have to be included only apply when multiple value elements are found.
# gRegorLove Fingers crossed, though, I think https://github.com/indieweb/php-mf2/pull/150 gets php-mf2 ready for release.
[kevinmarks] joined the channel
# 
[kevinmarks] The tricky date things are day of year dates eg 2018-256 for my birthday
# [kevinmarks] And yearless dates for birthdays where the year is hidden
# gRegorLove aaronpk, if you feel comfortable, I think we're good. You've been using it in action a lot more lately via Monocle.
# gRegorLove 👍
chrisaldrich joined the channel
# gRegorLove woo!
[asuh] and [chrisaldrich] joined the channel