#microformats 2018-07-25

2018-07-25 UTC
[jgmac1106], [ryan339] and [kevinmarks] joined the channel
#
@chrismessina ↩️ @dens @Foursquare @microformats Omg, we just switched places/purposes! (twitter.com/_/status/1021921350011772929)
[chrisaldrich] and snarfed joined the channel
#
snarfed hi all! hey aaronpk KartikPrabhu, i'm working on #11 on https://pin13.net/mf2/whitespace.html
#
snarfed it expects six spaces before "one", and there are six spaces inside the <pre>...
#
snarfed ...but there's *also* whitespace after "World", which i get should be collapsed, but i'm surprised that it should be removed entirely...?
#
snarfed as opposed to collapsed down to just one space, which would result in 6 + 1 == 7 spaces total efore "one"
#
KartikPrabhu I don't have any whitespace expectations so don't ask me :P
#
snarfed hah too late, that ship sailed when you accepted (even partial) maintainership of mf2py
#
aaronpk Oh gosh i haven't looked at these in a while
#
snarfed np!
#
KartikPrabhu sure. but I don't have expected results in my head. Those all look as arbitrary to me as any other expectation
#
snarfed i also found https://wiki.zegnat.net/media/textparsing.html useful
#
snarfed KartikPrabhu: if you want to maintain mf2py effectively, you'll need to get them into your head!
#
KartikPrabhu nope! just like I have gotten by without understanding VCP and just following spec
#
snarfed hah, sorry, no. mf2py has to implement mf2 parsing. i don't think you can do that without understanding mf2 parsing :P
#
KartikPrabhu I understand mf2 parsing spec. but have no expected results for whitespaces
#
snarfed mf2 parsing spec includes whitespace
#
snarfed via textContent
#
KartikPrabhu if you say 7 spaces or 6 spaces they are the same to me
#
KartikPrabhu I'll make sure mf2py does what people expect it to do
#
snarfed that's not what you said in the PR review :P
#
KartikPrabhu that is what I said. I said the test was modified and didn't match aaronpk's expected result
#
snarfed right. but again, to be an effective owner, you'll need to at least try to understand what your code does :P
#
snarfed true!
strugee joined the channel
#
KartikPrabhu I know what the code does, but am completely ambivalent to "Expected" results
#
snarfed ambivalent is fine! as long as you *understand* the expected results
#
snarfed anyway. between aaronpk and Zegnat, i'm sure we'll find a 6 vs 7 answer
#
KartikPrabhu I don't understand why they are "expected" like 7 vs 6 spaces
#
KartikPrabhu sure
#
snarfed yeah that's kind of a problem. but we can agree to disagree
#
snarfed i mean hell, i can understand a rationale for either way. we just need an authoritative answer
#
KartikPrabhu also, as you can see the "expected" breaks for <p> in mf2py are different from the whatwg thing
#
KartikPrabhu so yeah, I don't know what is expected
[dougbeal], [matpacker] and [kevinmarks] joined the channel
#
tantek__ lots of whitespace discussion!
#
KartikPrabhu more like any consensus on what is expected
#
KartikPrabhu right now it seems lke the only people with "expected" whitespacing are aaronpk and snarfed. Zegnat is following some algorithm from whatwg (though not in its entirity) and I have no expectations
#
gRegorLove snarfed: you mean the whitespace before <pre> should be collapsed to 1?
#
KartikPrabhu gRegorLove: yes I think that is snarfed's point
#
KartikPrabhu or expectation
#
gRegorLove My reading of Zegnat's algorithm gives "Hello World" for the first child (Text node), then for the <pre> it runs the algorithm on its contents, giving "\n one..."
#
gRegorLove Then append them, so 6 spaces looks right.
#
KartikPrabhu yeahit deviates from Zegnat's algorithm
#
KartikPrabhu mostly because it is trying to fix this https://github.com/microformats/mf2py/issues/112
#
Loqi [snarfed] #112 inner whitespace in p- values shouldn't be collapsed
#
KartikPrabhu but in the process breaks other stuff
#
gRegorLove what deviates?
#
KartikPrabhu snarfed's new algorithm
#
gRegorLove Ah
#
KartikPrabhu and hence expectation
#
gRegorLove I don't follow why he's expecting 7 spaces instead of 6
#
KartikPrabhu because to fix the issue 112 cited above, spaces between strings and HTML tags need to be collapsed to 1 and not 0
#
KartikPrabhu if phpmf2 is doing something magically correct let me know
#
gRegorLove Checking. It is giving the expected "foo bar" for snarfed's first example.
#
KartikPrabhu sure but does it pass the tests in https://pin13.net/mf2/whitespace.html apart from the one which is incorrect https://github.com/aaronpk/microformats-whitespace-tests/pull/1
#
gRegorLove I don't know. Probably. I was focusing on #11 since snarfed mentioned it, and now on the first example in #112.
#
KartikPrabhu #11 is the one which is the incorrect one but only with respect to the \n,
#
gRegorLove Here's php-mf2 parse with some debugging it gets "\n foo " for the first Text node
#
KartikPrabhu "a \n after the <pre> opening tag is neglected"
#
gRegorLove Oops, link: http://gregorlove.com/php-mf2/test.php?id=63
#
KartikPrabhu sure. that is also what snarfed's algorithm does
#
KartikPrabhu but that probably also puts 7 spaces after the #11 example instead of 6
#
KartikPrabhu unless phpmf2 is doing something really magical!
#
gRegorLove Here's #11 http://gregorlove.com/php-mf2/test.php?id=64
#
KartikPrabhu yeah that is completely wrong since it collapses spaces inside a <pre>
#
gRegorLove I guess that algorithm is missing <pre>
#
KartikPrabhu yeah Zegnat's algorithm does not say anything about <pre>
#
gRegorLove Personally I much prefer "Hello World one two three" for the p-name in this example
#
KartikPrabhu well good, we have more different "expectations" now :P
#
KartikPrabhu so everyone with some "expectation' can sort it out and let me know in the end
#
KartikPrabhu that is the problem with this whitespace thing. There is one set of expectations from aaronpk and there is an algorithm by Zegnat and they don't match. And others have some other expectation
#
gRegorLove I'm looking for the <pre> whitespace discussion but can't find it. I recall aaronpk talking about it. Maybe it was only in chat, not github
#
KartikPrabhu not sure
tantek joined the channel
#
gRegorLove Looks like some of the whitespace-tests were added after Zegnat's draft algorithm, so that explains the divergence. https://github.com/microformats/microformats2-parsing/issues/15#issuecomment-375595529
#
Loqi [Zegnat] It only took 2 months, but I have written a draft specification for handling `textContent` in the microformats context. The algorithm combines: * #17, always remove `<script>` and `<style>` elements, replace `<img>` elements, and strip leading an...
#
@jo ↩️ @dchatenay "microformats" (twitter.com/_/status/1021975214295404547)
#
KartikPrabhu gRegorLove: more examples if you want to document your expectations https://github.com/aaronpk/microformats-whitespace-tests/issues/4
#
Loqi [kartikprabhu] #4 whitespace between tags
#
gRegorLove I don't really have strong expectations, just thought that value looked better for a p-name
#
KartikPrabhu ok
#
gRegorLove Mostly trying to get up to speed on the topic
#
gRegorLove I'll check that issue
#
KartikPrabhu at this point the only way to speed up the topic would be collect test cases and get consensus on what is "Expected"
#
KartikPrabhu oh sorry "up to speed" not "speed up" :P
strugee, tantek__, [kevinmarks] and KevinMarks joined the channel
#
sknebel re <pre> tags, there's a good argument for adding newlines after <pre> too (if it doesn't already end with one), since browsers seem to by default treat it as display:block
strugee, jgmac1106, adactio, adactio_, KevinMarks, [jgmac1106], dmcweeney and [eddie] joined the channel
#
@cswordpress ↩️ Still working on adding microformats 2 support to #genesiswp child themes but I'm going to temporarily switch away because I don't want to wait for that to be done to start owning content and incorporating as many indieweb principles as possible on that website. (twitter.com/_/status/1022132264682106882)
KevinMarks, snarfed and KevinMarks_ joined the channel
#
sknebel snarfed: thoughts on the newline in front of <pre>?
#
snarfed sknebel: not really, no, sorry. i never felt strongly about it; the WHATWG rule kartik pointed to always seemed fine to me
#
sknebel that rule has nothing to do with my argument :/
#
sknebel it's about what happens to newlines inside the <pre>
#
sknebel I suggest treating <pre> like a new paragraph, like browsers do by default
#
snarfed the whatwg rule sure seems to apply to me: https://html.spec.whatwg.org/multipage/grouping-content.html#the-pre-element
#
snarfed "In the HTML syntax, a leading newline character immediately following the pre element start tag is stripped."
#
sknebel of course it applies
#
snarfed ok, then i don't follow
KartikPrabhu joined the channel
#
snarfed personally i'm more interested in the other q from last night: does the whitespace *before* <pre> get collapsed to one space or none?
#
snarfed but i don't feel strongly about that either
#
sknebel if you put "some text <p>content</p>" on a page, you get a newline for the <p>, and the whitespace after "text" gets removed cause there is a newline
#
sknebel in a browser, <pre> does the same
#
sknebel I ask for our code to do the same for <pre>, too
KevinMarks joined the channel
#
snarfed ahh i see, got it
#
snarfed sure, sounds godo
#
sknebel which also solves your question about the whitespace before it ;)
#
snarfed heh, not exactly, we'd still need to determine whether there would be a space before the newline or not
#
snarfed we could just say, no space
#
sknebel hm, I was under the impression space before newline gets always collapsed was clear, but I might be assuming too much
#
snarfed nah, could be!
snarfed left the channel
#
sknebel gonna wait for 1-2 more comments on the <pre> paragraph thing (aaronpk, opinions? since it is "your" whitespace repo?) and if they're positive prepare a few test cases. or should I just make a PR against the whitespace repo now, to make it clearer what I mean?
#
aaronpk I think my expectations boil down to how do browsers render it
jgmac1106 and [snarfed] joined the channel
#
KartikPrabhu aaronpk: that is misleading since browsers also take CSS into context, also the number of "new lines" for <p> from your expectations is different from the algorithm browsers are supposed to use (for example)
#
aaronpk ignoring css
#
aaronpk like lynx for example
#
sknebel KartikPrabhu: oh, there was a difference with <p>? where?
#
KartikPrabhu yeah trying to find the spec. I think now mf2py uses 1 newline before and 0 after but it is supposed to be 2 before or something
#
KartikPrabhu Zegnat: do you recall which spec has the number of "new lines" for "<p>" when getting textContent?
#
KartikPrabhu aaronpk: also I was wrong textContent is supposed to be CSS agnostic. innerText is not
[jsorge] and [cleverdevil] joined the channel
#
KartikPrabhu these specs are annoying to parse!
#
sknebel yeah. gotta admit, I've been playing with browser devtools and looking at what happens mostly recently
#
@ELLIOTTCABLE remember microformats? (twitter.com/_/status/1022150719581896704)
#
KartikPrabhu gives up on finding a reference for the <p> spacing
#
sknebel we'll find it again if it becomes important
#
aaronpk wow, 5 years later @elliottcable https://chat.indieweb.org/2013-06-22#t1371938182000000
#
Loqi [sandeepshetty] aaronpk: based on the tone of his comment I doubt he will buy anything we have to say :) plus from his own admission ("I don’t write a blog" - https://twitter.com/ELLIOTTCABLE/status/348537348151193600) he is looking more at "apps" than content
[jgmac1106], KevinMarks_, KevinMarks and tantek__ joined the channel
#
@cswordpress ↩️ You'll want a theme with microformats 2 support baked in. There's an accessibility ready one at https://github.com/pfefferle/ZenPress (twitter.com/_/status/1022162622932832258)
jackjamieson, gRegorLove and KartikPrabhu joined the channel; gRegorLove left the channel
#
gRegorLove Following https://dom.spec.whatwg.org/#dom-node-textcontent, I don't get 2 newlines before <p>. Matches my test in Chrome devtools
KevinMarks, [snarfed], [eddie], jackjami_, KartikPrabhu and [cleverdevil] joined the channel
#
@MarcoZehe @yellowled Aber hey, der Permalink ist jetzt da, das neue Theme ist also definitiv besser mit den Microformats. (twitter.com/_/status/1022187382089678848)
[kim_landwehr] joined the channel
#
tantek__ Ja, besser mit den Microformats
[eddie] and [pfefferle] joined the channel
#
[pfefferle] 😏
[jgmac1106] and [manton] joined the channel
#
sknebel edited /mirrors (+112) "add note about update policy" (view diff)
KartikPrabhu, bradenslen, [kevinmarks], [matpacker], tantek and jgmac1106 joined the channel
#
@rubygems indieweb-post_types (0.3.1): IndieWeb Post Type Discovery for Microformats 2 JSON. https://rubygems.org/gems/indieweb-post_types (twitter.com/_/status/1022240297718345729)
#
tantek interesting
[jgmac1106] joined the channel
#
@kickscondor Slurping up lots of HTML, so that I can get comment counts. But the microformats templates in use out there are all over the place. I can’t imagine the number of templates in use across the Indieweb. (twitter.com/_/status/1022242694855311365)
[cleverdevil] joined the channel
#
KartikPrabhu that is the point of microformats, you can use them in a variety of ways
#
sknebel KartikPrabhu, snarfed: example proposals for <pre> https://github.com/aaronpk/microformats-whitespace-tests/pull/5
#
Loqi [sknebel] #5 Add tests showing <pre> on same line as other content
#
KartikPrabhu sknebel: in test number 13 why is the space between </pre> and "And Hello Mars" collapsed?
#
sknebel Cause there's a newline, and that's what the browser does
#
KartikPrabhu ok
#
sknebel I don't think there's ever a reason to keep space around newlines except inside pre
#
KartikPrabhu ok at this point I'm just going to go with what people want
#
gRegorLove Zegnat's going to have fun catching up on all this when he's back :)
#
sknebel Heh yep
#
tantek KartikPrabhu: not sure about "what people want" - good technology, like science, is not democracy
#
KartikPrabhu tantek: if you have some source of how whitespace should work other than people's expectation then sure
#
KartikPrabhu you know like a theorem or experiment
#
tantek or you point to an existing dfn per principle of least invention / least surprise
#
KartikPrabhu yeah why me?
#
KartikPrabhu I don't consume anything yet that needs whitespace stuff
#
tantek even "people's expecations" can be measured better (by watching what they do) or poorly (by asking them and listening to them tell you)
#
KartikPrabhu yes, which is what is going on here
[snarfed] joined the channel
#
[snarfed] @kartikprabhu it's you (along with others) because you maintain an mf2 parser 😁
#
tantek which?
[matpacker] joined the channel
#
[snarfed] if you want to abdicate that responsibility, then yes, you wouldn't have to care nearly as much!
#
[snarfed] mf2py
#
KartikPrabhu ok, I wouldn't have done that if someone else had taken it up
#
KartikPrabhu ok, I am abdicating mf2py
#
[snarfed] hey, yup, the agony and ecstasy of open source
#
[snarfed] oh man really? sigh, if so it will be abandoned again 😢
#
KartikPrabhu yes
#
[snarfed] aww
#
KartikPrabhu there we go. officially done
#
[snarfed] how about a deal: we help you get 1.1.2 out with the current time of fixes, then you can ignore it for a while
#
KartikPrabhu naah it is done already
#
[snarfed] right now it has a bunch of good, pending bug fixes. it's a particularly harmful time to abandon it
#
gRegorLove :/ Well I hope you know your work on it has been very appreciated.
#
[snarfed] hopefully you'll at least find a successor?
#
KartikPrabhu no, if there was someone willing to take it up it would have been done a while ago instead of me
#
[snarfed] not necessarily true. never underestimate the power of asking
#
KartikPrabhu not my problem now
#
KartikPrabhu I have my fork for my site
#
[snarfed] wow.
#
[snarfed] please consider at least getting this release out before you abandon us!
#
tantek KartikPrabhu: I apologize I didn't mean to discourage you!
#
[snarfed] doesn't even need to settle open whitespace qs!
#
tantek right that ^^^
#
[snarfed] just merging the existing fixes would be enough
#
tantek KartikPrabhu: feel free to ignore all the whitespace bugs etc until people have a strong consensus based on examples and reasoning instead of "what people want" literally, just ignore those issues
#
tantek because fixing/merging all the others is so much more valuable
#
tantek snarfed have you tried out the 'alt' handling BTW?
#
gRegorLove ^that's what I've largely been doing until recently when curiosity got my attention, heh.
#
tantek I think we're still waiting to confirm that the prototype changes for doing 'alt' work as expected and satisfy your (snarfed) use-case
#
[snarfed] [tantek] yes, i believe the new alt support works for me
KevinMarks_, tantek__ and snarfed joined the channel
#
snarfed hey KartikPrabhu could you please add me to https://github.com/microformats and the mf2py pypi project? i'll get the release out myself if necessary. :(
#
KartikPrabhu don't think I can anymore
#
KartikPrabhu or even before
#
snarfed try and see?
#
snarfed if not that's ok
#
KartikPrabhu yeah tried. Can't, not owner or something
#
snarfed on both?
#
KartikPrabhu yup
#
tantek__ snarfed, just added you to the org I think
#
tantek__ does that help?
#
snarfed thanks. some, yes
#
snarfed hey KevinMarks_, KartikPrabhu has abdicated responsibility for mf2py. could you please add me as a pypi project owner?
#
snarfed tantek: i also need write access to the mf2py repo. any chance you can add me there? https://github.com/microformats/mf2py/settings/collaboration
[dougbeal] joined the channel
#
tantek__ I think so. just added both you and KartikPrabhu - not sure why / how Kartik wasn't there before?
#
KartikPrabhu I just left
#
tantek__ 😂
#
snarfed tantek: thanks! i still need *write access* to the repo
#
tantek__ snarfed: I'm looking at https://github.com/microformats/mf2py/settings/collaboration and it says you have "Write"
#
tantek__ along with Tom and Kyle
#
snarfed huh ok. still doesn't let me merge PRs. guess i'll give it a while
[cleverdevil] joined the channel
#
tantek__ snarfed, since the other two aren't around much, bumped you to "Admin" on there
#
snarfed hah sure
#
tantek__ we should have at least one admin per project that isn't just a microformats admin
#
Loqi yea!