[cleverdevil], KartikPrabhu, davidmead, nitot and gRegorLove joined the channel
# 07:05 Loqi [sknebel] #126 Fails to recognize timezone in extra value-class-pattern-element
[colinwalker] joined the channel
# 07:07 KartikPrabhu gRegorLove: my first reaction is that it should be left to the consumer and not the parser
# 07:52 Loqi Ok, I'll tell them that when I see them next
# 07:56 Loqi Ok, I'll tell them that when I see them next
nitot, adactio, [colinwalker], rodolfojcj and ben_thatmustbeme joined the channel
# 12:56 Loqi Ok, I'll tell them that when I see them next
# 13:30 sknebel !tell gRegorLove: nevermind, it only talks about implied timezones on a different section of the page, so this really could use clarification (*not* implying timezones but implying dates (which php-mf2 already does) seems really odd)
# 13:30 Loqi Ok, I'll tell them that when I see them next
# 13:52 Zegnat I would say everything that is defined on dt-start carries over to dt-end, including tz info? (Is that what this is about?)
# 13:58 Zegnat Isn’t this up to consumers though? Or do mf parsers return normalised timestamps? I was under the impression that parsers should only return what was actually on the page (e.g. "end" would get value "21:00" not "2017-05-31T21:00+02:00")
# 14:01 Zegnat Oh wow I somehow glanced over the very first paragraph.
# 14:01 sknebel happens. it took me ages to find and realize it applies as well
# 14:06 Zegnat So dt-* values with only a time only happen when there isn’t a single dt-* including a date on the entire page?
# 14:06 Zegnat Or what does “in that microformat” mean? That it is scoped to the current h-*?
Garbee_ joined the channel
# 14:08 Zegnat Oh, hmm, and it only does applied date when vcp is used. So if I do <abbr class="dt-end" title="21:00">9 o’clock</abbr> the mf2 parser must return only "21:00".
# 14:14 Zegnat <abbr class="dt-end" title="21:00">9 o’clock</abbr> = ["end"=>"21:00"]
# 14:14 Zegnat <span class="dt-end"><abbr class="value" title="21:00">9 o’clock</abbr></span> = ["end"=>"2017-05-31T19:00"]
# 14:15 Zegnat Actually, your current example wouldn’t get an implied date on the dt-end either, as it isn’t using vcp for dt-end
# 14:17 sknebel and I can't turn it into a vcp without adding another nested element
# 14:18 sknebel sounds like that's worth a more general issue, before I start complaining to more parsers about what they do
# 14:18 Loqi [Tantek Çelik] microformats2 parsing specification
# 14:19 Zegnat Not so much a discussion of vcp (which is pretty clear) but on implied dates. Maybe implied dates should be removed from vcp and be moved to dt-* parsing?
# 14:20 sknebel maybe, there'd be easier to find there as well...
# 14:22 Loqi [tantek] #4 vcp: imply dates should also imply tz if needed
Garbee joined the channel
# 14:30 Zegnat I am pondering how to writup an issue on implied dates in general. I can come up with examples, but not really a clear course of action
# 14:47 sknebel make a gist or etherpad and let's discuss/interate there before filing it?
# 15:05 Zegnat Probably a good idea. My evening is pretty full though, so not sure I have time to sit down and look atit
# 15:47 sknebel wasn't entirely sure about the meaning, and twitter translate button made it worse :P
# 15:47 Zegnat They were asking how much support for v2 there is, compared to v1. So told them indieweb.org and W3 standards like Micropub are all v2, as well as parsers (with microformats.io link). If they do find a specific usecase for v1 I asked them to get in touch
# 15:48 Loqi zegnat has 2 karma in this channel (54 overall)
tantek joined the channel
# 16:09 sknebel Zegnat: seems like various parsers do the implicit date when they shouldn't
# 16:11 Zegnat Would have been easier if all of them just did it and we could write it up as a defacto standard
# 16:21 sknebel 3 of those have the timezone bug, the JS one does parse the timezone correctly and implies it onto dt-end
# 16:45 Loqi [@ankedesign] @martijnvdven Ha, dank! Heb v2 gebruikt, maar twijfelde omdat bv o.a. de Google checker V2 niet herkent.
tantek, [shaners] and gRegorLove joined the channel
# 17:31 gRegorLove Not sure about format checkers, but search still uses hReview
# 17:31 Loqi gRegorLove: sknebel left you a message 4 hours, 1 minute ago: nevermind, it only talks about implied timezones on a different section of the page, so this really could use clarification (*not* implying timezones but implying dates (which php-mf2 already does) seems really odd)
# 17:32 gRegorLove I've got the VCP timezone fix working locally. Will run some more tests and PR today or tomorrow.
# 17:32 Loqi gregorlove has 8 karma in this channel (138 overall)
# 17:32 Loqi [tantek] #4 vcp: imply dates should also imply tz if needed
# 17:33 Loqi tantek: KartikPrabhu left you a message 9 hours, 37 minutes ago: spam! ^
# 17:34 tantek gRegorLove: wow they started putting "." in words to avoid the regexes!!!
# 17:42 Loqi [tantek] #4 vcp: imply dates should also imply tz if needed
# 17:42 Loqi [Zegnat] Oh, hmm, and it only does applied date when vcp is used. So if I do <abbr class="dt-end" title="21:00">9 o’clock</abbr> the mf2 parser must return only "21:00".
# 17:42 gRegorLove tantek: I did not; raised the question about it last night; hadn't seen the issue yet.
# 17:45 Loqi [sknebel] #126 Fails to recognize timezone in extra value-class-pattern-element
# 17:45 tantek non-VCP dt-* implying could be both useful and dangerous (in terms of changing the content that's out there)
# 17:46 tantek that seems to imply that it's intended in the content then, I doubt that many parsers would do extra work to misinterpret content like that
# 17:46 tantek sknebel: are you writing a parser issue to suggest adopting it as a parser change?
# 17:47 gRegorLove on indieweb.org/events we pretty commonly use non-VCP dates with the date part implied in the dt-end.
# 17:47 sknebel no, in microformats-parsing-issues to ask if it's supposed to be handled as bug in the parsers or maybe changed in mf2
# 17:47 Zegnat I would guess it is seen as “intended” because very few end times use vcp, yet dt-end on events is where people expect the implied dates.
# 17:50 Loqi [sknebel] #8 imply dates also outside vcp?
# 17:53 tantek gRegorLove: html-lang and html-id to avoid confusing them with a possible actual property p-lang or p-id (which we don't have but might / could, especially from a vocabulary agnostic parser perspective)
# 17:57 gRegorLove tantek: Makes sense. Yeah, I'll add it there. That's released in php-mf2 now, btw, behind a feature flag.
tantek, AngeloGladding and Garbee joined the channel
# 19:58 gRegorLove tantek: Looking at the lang parsing again, the "html-lang" key doesn't show up in properties, only within e-* objects and at the same level as "type" and "properties", so maybe the prefix isn't necessary?
# 20:03 gRegorLove And I think if it was ever expanded to allow parsing on individual properties, like p-name, we'd need to come up with some other solution
# 20:04 aaronpk yeah, it's a little unfortunate it doesn't translate nicely for things like p-name
# 20:04 aaronpk i think the closest you'd get is changing the value of the name property to an object like how e-content is parsed
# 20:05 gRegorLove Yeah, so I don't think 'lang' and 'id' are at much risk of collision in the future
# 20:06 aaronpk yeah it will never exist alongside vocab properties
[kevinmarks] joined the channel
tantek, pfefferle, [colinwalker], cyr_ and [cleverdevil] joined the channel
Garbee and [eddie] joined the channel