#microformats 2020-06-29

2020-06-29 UTC
[Beto] and maxwelljoslyn joined the channel
#
[tantek] oh dear we need to update the creators 🙈
#
[KevinMarks] a project for next camp
#
[KevinMarks] get Jordan to check with indiewebify.me
#
gRegorLove https://github.com/microformats/generators
#
[tantek] oh good I did check them in
#
gRegorLove I'll update on the indieweb wiki
#
[tantek] but seriously don't bother trying to understand the JS there
#
[tantek] asking people to learn an obsolete JS framework seems harmful
#
[schmarty] heehee
#
gRegorLove If I tried to update I'd just use vanilla js
#
[tantek] I can never tell if that means "plain" js or a framework called "vanilla" 😛
#
[schmarty] the vanilla JS framework is a beautiful joke
#
gRegorLove artisinal, small-batch js
#
[KevinMarks] is templates the framework or jquery?
#
[KevinMarks] jquery transaltions: http://youmightnotneedjquery.com/
#
gRegorLove yep, love that site
#
[schmarty] what is Vanilla JS?
#
gRegorLove <- indieweb chat?
[Paulo_Pinto] joined the channel
#
KartikPrabhu [tantek] if all it is doing is generating HTML from the form fields, it is the same as the little notes post UI i demoed
#
KartikPrabhu so I can probably write it in plain old JS
#
gRegorLove Yeah, sounds like a familiar project... XD
#
[KevinMarks] string concatenation is a template language too
#
[KevinMarks] I fine nunjucks really handy, but it may be overkill
#
KartikPrabhu JS has built-in string templating now! which is what I was using
#
jacky ^
#
jacky that'll be a lot less resources to use
[scottgruber] and [mJordan] joined the channel
#
[tantek] yeah totally. native JS has far more features than when we built these creators back in 2005
[chrisaldrich], gRegorLove, beko, [georgenancejr], [christopherche], opengoody, [fluffy], [KevinMarks], [jgmac1106], [grantcodes] and [Murray] joined the channel
#
ben_thatmustbeme I remember this coming up in the past, what are thoughts on multiple parsers per language?
#
[KevinMarks] that depends on other packaging too - client-side js parser vs node one may be different domains for example; I can imagine other constraints within frameworks eg making it fit an api idiom
#
[KevinMarks] wasn't there some variation on the PHP side?
#
[KevinMarks] if we can get the tests and test running smoothly between languages, presumably hooking them up to a different substrate makes sense. Another reason might be to switch to a fully html5+ compliant parser rather than an xml-biased one maybe
#
sknebel if you're asking "is it a good idea", I'd say "if it fixes something that won't be fixed in the existing one(s), maybe". Or if it replaces one that's effectively abandoned
#
sknebel i.e. if I remember right someone started one as a native PHP module. that'd be harder to install, but would hopefully be noticably faster
#
aaronpk The first ruby one didn't return a data structure and instead had magic methods to return data and it was pretty annoying so I was glad to see it rewritten
#
aaronpk It did the very ruby thing of accessing a property with the plural name would return an array, eg doc.hcards vs doc.hcard
#
aaronpk item.url vs item.urls
#
aaronpk im sure you can imagine the edge cases where this fails awkwardly :-)
marinin[m] joined the channel
#
ben_thatmustbeme yeah, i know, i rewrote the parser
#
ben_thatmustbeme in ruby
#
ben_thatmustbeme but now i'm turning my eye toward libs in php
#
ben_thatmustbeme and going to rewrite pretty much any that don't fit my needs, which may be all of them
#
ben_thatmustbeme and at that point i started to look at the parser and it seems like a lot of the mf1 tests fail (and are disabled now) and I'm wondering if i should just redesign it in a way that makes sense to me
#
ben_thatmustbeme i sort of like the plurality of design type idea, but yeah, when it comes to parsers, its a harder sell
#
ben_thatmustbeme it would also be a different case if there weren't existing failing tests and tickets still open for years
[snarfed] joined the channel
#
[KevinMarks] there was some effort to clarify the tests repository - have a look at the checkins and issues
#
[KevinMarks] if we can converge on using the same test data across languages that seems like progress (historically we had a lot of separate tests as well as fixtures)
jamietanna[m] joined the channel
#
aaronpk ben_thatmustbeme++ for rewriting the ruby parser :)
#
Loqi ben_thatmustbeme has 1 karma in this channel over the last year (2 in all channels)
#
ben_thatmustbeme well thats what the current ruby parser is haha
#
ben_thatmustbeme i would honestly just follow the design i did for that, as it was very different from what is there now
KartikPrabhu, [fluffy], [jgmac1106], omz13, [hibs] and [tantek] joined the channel
#
GWG I am confused...why do we need another PHP parser?
#
GWG I'm just missing the impetus
[chrisaldrich] and gRegorLove joined the channel
#
ben_thatmustbeme just because there is only one and it has holes in it. its not necessarily needed, this is how my brain works though. Ooh, a thing thats not perfect... i could completely rewrite it
#
aaronpk could you start by documenting the holes as issues on the current one?
#
[tantek] filingissues++
#
Loqi filingissues has 1 karma over the last year
#
[tantek] ben_thatmustbeme, I can totally relate to "this is how my brain works though" and would definitely encourage you to start with capturing the issues (helps the folks maintaining the existing PHP parser, there's quite a few of them / us - I help with minor reviews lol)
#
[tantek] and that being said, designing / writing a new library for something like that from scratch I can totally understand, often you can make totally different implementation decisions that can lead to faster or smaller code, or fewer dependencies etc.
#
ben_thatmustbeme well i think they are pretty well documented, plus there are already tests in place
#
ben_thatmustbeme i'll try to skip over it and put my focus on other bits first
#
[tantek] ah, existing issues. got it
#
GWG ben_thatmustbeme: I am curious what you come up with regardless
[christopherche], [KevinMarks], indy, [JuJu], [tw2113] and [schmarty] joined the channel
#
cjw6k i did the mf2 php-extension, but a career change led to it getting dropped. It's still here and I have been using it lately. https://mf2.w6k.ca/
#
cjw6k I may pick it up again before too long, we'll see. No immediate plans.
#
cjw6k I think I'll try to get it into PECL if I get it into a better state
#
gRegorLove I have had that temptation about php-mf2 as well even while working on it, hah
#
gRegorLove btw, if someone can merge this I think that's the last item currently for the 0.5.0 milestone https://github.com/microformats/php-mf2/pull/201
#
Loqi [gRegorLove] #201 Fix vevent location backcompat
#
cjw6k php-mf2 is very complex
#
cjw6k as in cyclomatic- and npath-