#microformats 2018-03-20

2018-03-20 UTC
[snarfed], [kevinmarks] and [miklb] joined the channel
#
[kevinmarks] Where this gets practical is in the tests repo, where you should not rely on order when comparing the dictionary like things
#
tantek either way the spec needs to be explicit
#
[kevinmarks] Yes
chrisaldrich, tantek and [miklb] joined the channel
#
KartikPrabhu !tell Zegnat: updated mf2py with "ordering" stuff. now "items" and "rels" should be deduped and sorted alphabetically. also up on my site version
#
Loqi Ok, I'll tell them that when I see them next
tantek joined the channel
#
@steve_e @alexismadrigal @KatzOnEarth If only people had used microformats for publish dates. (twitter.com/_/status/975990074016968704)
nitot and KartikPrabhu joined the channel
#
Zegnat Got my tea, time to do issue writing
#
Loqi Zegnat: KartikPrabhu left you a message 2 hours, 46 minutes ago: updated mf2py with "ordering" stuff. now "items" and "rels" should be deduped and sorted alphabetically. also up on my site version
#
tantek I hope that's a typo. because "items" should NOT be deduped or sorted alphabetcially.
#
Zegnat It doesn’t look like mf2py is going anything to items.
#
Zegnat bleeding edge mf2py does alphabetically sort type array (as PHP already did) and dedupe its values (as is logical for a "set")
tantek joined the channel
#
Zegnat !tell gRegorLove please see if this clears up your questions about yesterday’s discussion: https://github.com/microformats/microformats2-parsing/issues/29
#
Loqi Ok, I'll tell them that when I see them next
#
Loqi [Zegnat] #29 Define the order of items any time an array is used in the parsed output.
[kevinmarks] joined the channel
#
@sbillard @VictorLerat Renvoyer de la popularité vers les niveaux supérieurs, afficher le chemin dans les SERPS avec les microformats adequats, et permettre une navigation plus aisée. En haut de page par contre, je vois pas l'intérêt du bas de page. (twitter.com/_/status/976029745874907136)
#
Zegnat Separate issue for deduping type and rels: https://github.com/microformats/microformats2-parsing/issues/30
#
Loqi [Zegnat] #30 Arrays type and rels should not contain duplicate items.
nitot joined the channel
#
Zegnat If people could take another look at https://github.com/microformats/microformats2-parsing/issues/28 and either come in with an objection or just +1 it, I think we can get that one closed relatively quickly?
#
Zegnat How is intent to implement on that one? Guessing gRegorLove wants it in PHP as he opened it. I see the bleeding edge mf2py supports numbers already.
#
Loqi [gRegorLove] #28 Allow numbers in vendor prefixes
tantek, nitot, barpthewire, [mrkrndvs], [kevinmarks], Garbee, [pfefferle], [miklb], webchat140, nitot_, TallTed, brandon and KartikPrabhu joined the channel
#
brandon Hi there. Wondering, how might I format savings text. Specifically, text that looks like “Save $1,200.00 if you buy the complete package”?
#
KartikPrabhu brandon: not really a microformats question
#
KartikPrabhu Zegnat: oops I did mean "type" in the tell I left you
#
KartikPrabhu also mf2py does not do any validation for "h-*" and properties yet. I am waiting for someone to write that regex
#
Zegnat KartikPrabhu, I kept the issues pretty general in the end, so no problem. Didn’t cite a whole lot of stuff. Feel free to expand on them or ask questions if I wasn’t clear on something
#
Zegnat I’d write the regex for you, but I do not know how to apply the regex in Python :P
#
KartikPrabhu no the issues are pretty clear IMO
#
KartikPrabhu Zegnat++
#
Loqi zegnat has 12 karma in this channel (181 overall)
#
KartikPrabhu Zegnat: I can do that bit :) you should put the regex on that issue!
#
Zegnat (h|p|u|dt|e)-([a-z0-9]+-)?[a-z-]+
#
Zegnat should probably do the trick
#
KartikPrabhu hmm ok will test thanks!
#
Zegnat I have to update my JSON schema with that too, so had to write it anyway
#
KartikPrabhu haha
#
Loqi hehe
#
Zegnat Basically says: "match any of these 5 property prefixes, then a literal -, then possibly any combination of letters and digits followed by another literal -, then one or more letters or literal -.
#
Zegnat Which I think covers it all
#
sknebel h------
#
Loqi h has -1 karma
#
sknebel lol
#
sknebel h++
#
KartikPrabhu h++
#
Loqi h has 0 karma
#
Zegnat sknebel, yeah, that’s a valid name per mf2
#
Zegnat Unless I have missed something in the spec
nitot joined the channel
#
sknebel you're currently working to change whats valid, so consider now if that should be the case
#
Zegnat I think that should be a separate issue. The current issue is really only for allowing numerals in prefixes
#
@chrismessina Life, gamified. (@ Starbucks in Mountain View, CA w/ @microformats) https://www.swarmapp.com/c/8ZYfGVtH4Wj https://pbs.twimg.com/media/DYvvn5iXUAAHcK3.jpg (twitter.com/_/status/976137215490297856)
#
@zacharyparadis @chrismessina @microformats Stars, shields, free stuff: https://en.wikipedia.org/wiki/Green_Shield_Stamps (twitter.com/_/status/976138137897521154)
#
KartikPrabhu Zegnat++
#
Zegnat Is that a sign that the regex worked?
#
KartikPrabhu yes!
#
KartikPrabhu writing tests now
#
KartikPrabhu where is Loqi?
#
Zegnat Maybe to soon after your previous ++ to me?
#
KartikPrabhu Zegnat++ for class names regex
#
KartikPrabhu boo!
#
KartikPrabhu well you got the point ;)
#
sknebel Zegnat++ ?
#
Loqi zegnat has 13 karma in this channel (182 overall)
#
Zegnat Haha
#
sknebel see, the bots like me :)
#
Loqi rofl
#
sknebel gives Loqi a cookie
#
Loqi gives back the cookie
#
Zegnat https://github.com/microformats/microformats2-parsing/issues/31 - filed, sknebel
#
Loqi [Zegnat] #31 Should a consecutive run of dashes (-) be allowed in root and property class names?
#
Zegnat Hmm. Way too many of those open issues are on my name
#
sknebel :)
#
KartikPrabhu Zegnat: here you go https://kartikprabhu.com/connection/mfparser :) with the names check
#
Zegnat Initial testing seems like it works perfectly KartikPrabhu :D
#
Zegnat KartikPrabhu++ for all the Python wrangling
#
Loqi kartikprabhu has 17 karma in this channel (184 overall)
#
KartikPrabhu :D
#
KartikPrabhu there wasn't much wrangling to do in this one. Just replaced the starts with "p-" with a regex check :D
#
Zegnat went looking for the code in microformats/mf2py
#
Zegnat Is this in a private fork?
#
KartikPrabhu Zegnat: https://github.com/kartikprabhu/mf2py/blob/experimental/mf2py/mf2_classes.py
#
KartikPrabhu yeah because super experimental
#
Zegnat Those regexes look right, for as far as I can read Python
#
KartikPrabhu yeah did some prior testing in console before adding them in
#
KartikPrabhu I am basically keeping my master branch to push to the main repo so all new stuff goes into my experimental
#
KartikPrabhu which then goes to my site so I can spot if it breaks in the wild
#
Zegnat selfdogfooding one-oh-one there KartikPrabhu, nice!
#
KartikPrabhu yup I did find a few bugs that way :)
[eddie], tantek and [cleverdevil] joined the channel
#
Zegnat KartikPrabhu, if you want to test with blocking consecutive dashes (cf. #31): (h|p|u|dt|e)-([a-z0-9]+-)?[a-z]+(-[a-z]+)*
#
KartikPrabhu are those not allowed?
#
sknebel I suggested we might not want to allow them ( https://github.com/microformats/microformats2-parsing/issues/31 )
#
Loqi [Zegnat] #31 Should a consecutive run of dashes (-) be allowed in root and property class names?
#
sknebel (and Zegnat was kind enough to file an issue for it)
#
KartikPrabhu Zegnat: yeah that also works to eliminate that case
#
Zegnat Oh, my regex actually also disallows trailing dashes ... didn’t set out to do that, but I am not against that
#
Zegnat That is, I really treated dashes (hyphens. whatever.) as delimiters. Hmmm..
#
KartikPrabhu yeah it does
#
KartikPrabhu I think that is good
#
KartikPrabhu I'll update mf2py in a bit
#
KartikPrabhu so the new rule is basically treat "-" as strict delimiters
#
Zegnat I am not sure what I think of that yet. It was just how I inadvertently created the regex...
#
Zegnat Adding to the issue now with further thoughts
#
KartikPrabhu I support not allowing "-" in the end
#
Zegnat https://github.com/microformats/microformats2-parsing/issues/31 feel free to leave thoughts and/or intent to implement, KartikPrabhu :)
#
Loqi [Zegnat] #31 Should a consecutive run of dashes (-) be allowed in root and property class names?
[Vanessa], [chrisaldrich], [eddie], KartikPrabhu and [gerwitz] joined the channel
#
KartikPrabhu Zegnat: done and added to mf2py
#
gRegorLove KartikPrabhu, is that live on https://kartikprabhu.com/connection/mfparser?
#
Loqi gRegorLove: Zegnat left you a message 10 hours, 34 minutes ago: please see if this clears up your questions about yesterday’s discussion: https://github.com/microformats/microformats2-parsing/issues/29
#
KartikPrabhu yes
#
KartikPrabhu let me know if you find some gaps
#
KartikPrabhu gRegorLove: here is the code if it helps for php-mf2 https://github.com/kartikprabhu/mf2py/blob/691475feb8c71b87f4c82071395db9c78f494748/mf2py/mf2_classes.py#L7-L16
#
gRegorLove I think I'm +1 on that, was just about to comment on the issue. If there's any consecutive dashes it won't be parsed, right? So `p-acme--thing` would be excluded.
#
KartikPrabhu yes
#
gRegorLove Sounds good for leading and trailing dashes for sure. Can't really think of a reason to use multiples in the middle, but if they were allowed in the middle that doesn't seem like a problem.
#
KartikPrabhu I am sure Zegnat will give us the correct regex if the rules change :P
#
Zegnat Ha, sure
#
gRegorLove Commented: https://github.com/microformats/microformats2-parsing/issues/31#issuecomment-374738226
#
Loqi [gRegorLove] +1 for not parsing with leading or trailing multiple dashes. Neutral for multiple dashes in the middle, though I could easily be convinced of that for simplicity. Contrived example: if someone published `u-in-reply--to` would it be better debuggin...
#
gRegorLove KartikPrabhu: Heh, I was confused when I highlighted the text in your parser results what the dot was that appeared. Guess it's your fragmention thing. :)
#
KartikPrabhu yeah :P
#
Zegnat Thanks for the comment gRegorLove :D bonuses for keeping the ball rolling
chrisaldrich, tantek and nitot joined the channel
#
KartikPrabhu tantek: thoughts https://github.com/microformats/microformats2-parsing/issues/31 ?
#
Loqi [Zegnat] #31 Should a consecutive run of dashes (-) be allowed in root and property class names?
#
KartikPrabhu getting another opinion before updating spec
#
tantek that is bizarre that it was even allowed
#
tantek checks the spec
#
KartikPrabhu spec only says allow "a-z" and "-"
#
KartikPrabhu if that looks good we can update spec for numbers in vendor prefixes and this hyphen stuff
#
tantek are there tests for rejecting bad props?
#
tantek and impls that pass those tests?
#
KartikPrabhu tantek: I have them in mf2py
#
tantek or rather at least one?
#
tantek ok
#
tantek did we get additional implementer consensus on the issues?
#
KartikPrabhu the experimental version not the main one
#
tantek also what do people think of rel=archived?
#
KartikPrabhu gregorlove is for it
#
tantek http://microformats.org/wiki/index.php?title=existing-rel-values&curid=4255&diff=66721&oldid=66690&rcid=103590
#
Loqi existing rel values
#
tantek https://sitemorse.com/rel-archived/
#
KartikPrabhu not sure why it is useful as I don't use it
#
tantek the description sounds reasonable, I can see using it, and know of sites that could use it
#
tantek I don't use it yet, but could see using it for some of my content I don't want to update, e.g. XHTML 1.x examples
#
KartikPrabhu this is not a parsing question right?
#
KartikPrabhu no parser checks for valid rels afaik
#
tantek no separate thread
#
tantek it's a question about adding the value to the rel registry
#
KartikPrabhu ok
#
KartikPrabhu yeah if it is in use with examples then looks fine
#
KartikPrabhu tantek: class names format test from mf2py https://github.com/kartikprabhu/mf2py/blob/experimental/test/examples/class_names_format.html
#
tantek there are no documented examples, just a claim in the recent edit that there are examples
#
tantek since the person that added it did actually spend the time to come back and re-add it, and at least claim there are examples, I'm willing to accept that there may be non-obvious examples we just don't know about
#
tantek but figured I'd ask folks's opinions here
#
gRegorLove tantek, KartikPrabhu: I'm generally in favor of the dashes change, neutral on consecutive dashes in the middle of a root/property name. See my comment on the issue.
#
tantek the dashes change is not a change in intent btw
#
gRegorLove Re: rel-archived, makes sense when you're linking to the archived URL. Makes me wonder how you'd indicate on the URL itself that it's archived.
#
KartikPrabhu tantek: yes I figured those rules were missed in the original writing
#
tantek I see that the spec was excessively permissive though if you read it literally
#
tantek gRegorLove: it's possible you can't (or don't want to) change the URL itself
#
tantek because it's archived!
#
tantek of something from the past, which at that point in the past, was likely "current"
#
tantek so presumably you want to keep it "archived" as-is, without changing it
#
gRegorLove Hm
#
tantek that's why it made sense to me
#
gRegorLove But you would want to go back to some other page(s) that link to it and add rel=archived?
#
tantek yeah, nav pages etc.
#
tantek this makes sense with old/obsolete test cases for example
#
tantek especially with versioned URLs
#
tantek where you don't want to change the URL and break permalinks out there
#
tantek nor do you want to change the page content itself (because it's a test)
#
gRegorLove Interesting. Sounds like more work that way, depending on # of links. I don't have a proposal for what to add to an individual page to indicate it's archived though, either.
#
tantek best you could do in addition would be maybe some additional HTTP headers
#
tantek right, that's likely more work
#
tantek whereas you may just have one nav pages pointing to a bunch of old pages
[cleverdevil] joined the channel
#
KartikPrabhu tantek: if you can add your thoughts/agreement to the dashes issue we should be able to update the spec
#
tantek gRegorLove: which parser are you maintaining again?
#
KartikPrabhu php-mf2
#
tantek cool
#
tantek KartikPrabhu: yes I'll add my opinion and try to knock out some minimal spec edits for those two
#
KartikPrabhu there is a suggested edit by Zegnat I think
#
KartikPrabhu oh woops there is a suggestion without the "dashes" thing https://github.com/microformats/microformats2-parsing/issues/28#issuecomment-374536421
#
Loqi [Zegnat] I am :+1: on allowing digits. I do wonder about including vendor prefixes, as there is no strict rule for recognising them as I understand it? In `p3k-drank` the `p3k` is a vendor prefix. But in `in-reply-to` the `in` isn’t. CSS doesn’t have t...
#
tantek KartikPrabhu: don't worry about editing specifics, I can come up with those
#
tantek remind me which issues we think we have community+parserdev consensus + prototyping?
#
tantek #31 and ... ?
#
KartikPrabhu #28 #29 #30 #31
#
KartikPrabhu gRegorLove: #29 and#30 need your input :)
#
tantek ok I'll write replies to those now
#
KartikPrabhu :thumbsup:
#
tantek checks http://microformats.org/wiki/microformats2#Background to see if he captured any of these details in original brainstorming / design
#
Loqi microformats2
#
gRegorLove I +1'd 29, no comments currently
#
gRegorLove +1'd #30 now as well
#
KartikPrabhu nice!
#
gRegorLove #27 is probably easy too. In case you don't have enough on your plate ;)
#
gRegorLove Suggesting we add some ordinal date examples to the VCP examples.
#
KartikPrabhu should add ordinal dates to mf2py
#
tantek :)
#
tantek while I'm reviewing #27-31 and commenting, I KartikPrabhu,Zegnat,gRegorLove take a look at http://microformats.org/wiki/microformats2-brainstorming#hReview_item_backward_compatibility in case you hadn't found it already
#
Loqi microformats2-brainstorming
#
tantek that page likely ought to also be processed into gh issues for consideration
#
tantek or we can postpone dealing with all of them til the mfdev meetup at IWS
#
gRegorLove ooh
#
KartikPrabhu i think gRegorLove was doing the "analyse examples" part
#
gRegorLove php-mf2 is doing that. That's been captured in the h-review parser compat section.
#
gRegorLove I believe it's actually the inverse of that I have not found yet -- hReview with fn, url, or photo properties that *don't* have an intervening 'item'
#
gRegorLove http://microformats.org/wiki/h-review##item+(without+vcard,+vevent,+hproduct)
#
Loqi h-review
#
gRegorLove http://microformats.org/wiki/hreview-v0.2 has fn, url, and photo as properties of 'item'. Was it early hReview 0.1 that didn't have 'item'?
#
Loqi hReview 0.2
#
tantek maybe
#
tantek a-ha found it! http://microformats.org/wiki/naming#Some_Details
#
Loqi Naming Principles
#
tantek sorry there was no way y'all were going to find that
#
tantek but just to illustrate original intent. I'll put in my comment
#
tantek that page needs some updating for mf2 but is largely pretty good still
benwerd joined the channel
#
tantek welcome benwerd !
#
tantek Zegnat++ for filing such good issues and follow-up comments
#
Loqi zegnat has 14 karma in this channel (183 overall)
#
tantek KartikPrabhu++ same
#
Loqi kartikprabhu has 18 karma in this channel (185 overall)
#
tantek gRegorLove++ same
#
Loqi gregorlove has 24 karma in this channel (224 overall)
#
tantek ok commented on 28-31, those are all related
#
tantek #27 is separate enough to consider separately (as a separate edit)
#
tantek someone else want to verify (with comments on #28-#31) that we have sufficient consensus to do an edit and cc @tantek so I know you've done the verification?
#
tantek will check back in a bit
#
KartikPrabhu done