#microformats 2022-12-10

2022-12-10 UTC
GWG-, GWG, [Michelle_Moore], gRegor, gRegorLove_, tiim, ur5us and barnaby joined the channel; btrem left the channel
#
IWDiscordRelay <j​acky#7226> hm is `--01-20` valid for a Microformats date value?
#
IWDiscordRelay <j​acky#7226> I'm noticing that the Rust parser thinks it isn't and it passes the mf2 test suite of values
#
IWDiscordRelay <j​acky#7226> I know of a few failures in parsing it has
#
IWDiscordRelay <j​acky#7226> namely around datetime values (expected: "2009-06-26T19:00+08:00", actual: "2009-06-26T19:00+0800"; from the test suite, just ran it)
#
IWDiscordRelay <j​acky#7226> and concat'ing temporal values (expected: "2009-06-26 22:00-0600", actual: " 22:00-0600")
#
barnaby jacky, does the rust parser try to parse all dt- parsed properties and return them to the user as some sort of rust datetime structure rather than a string?
#
barnaby the only part of the mf2 parsing spec which makes any attempt to understand datetime values is the value-class datetime rules https://microformats.org/wiki/value-class-pattern#Date_and_time_values
#
[tantek] jacky, yes that's a valid microformats date value for e.g. a bday where the person doesn't want to publish the year
#
barnaby otherwise, parsers treat values as opaque strings and it’s up to the consumer to validate them
#
[KevinMarks] The use case for that is birthday without year. It was deprecated from ISO, but is still useful
#
[KevinMarks] See https://html.spec.whatwg.org/dev/text-level-semantics.html#datetime-value
#
aaronpk weirdly that format got removed from the 2004 version of ISO8601 but it is allowed in the 2000 version https://en.wikipedia.org/wiki/ISO_8601#Calendar_dates
#
barnaby it’d be interesting to see a review of working software which publishes/consumes the [--]-MM-DD format
#
barnaby or even datetime libraries which can represent it
#
barnaby [--]MM-DD, rather
#
Loqi [ has -1 karma in this channel over the last year (-2 in all channels)
#
barnaby oops
#
barnaby [++
#
barnaby apple calendar displays birthdays from contacts, so maybe they’re internally represented like that
GWG-, [jeremycherfas] and GWG joined the channel
#
IWDiscordRelay <j​acky#7226> Yeah there's a set of Rust objects to represent the time in the ways it can be accepted ("strict" validation, if you may). I'll add a test on my end but do people think adding a case to catch this in the test suite would be good?
GWG-, tiagorangel11, [manton], ur5us, tiim, barnaby, GWG and gRegor joined the channel
#
[KevinMarks] In the mf2 test suite? Yes, make it a birthday so the yearless date concept is clear.
GWG joined the channel