#microformats 2018-02-17

2018-02-17 UTC
PeterFD, chrisaldrich, tantek, globbot, [mrkrndvs], [colinwalker], [kevinmarks], sknebel, Zegnet and KartikPrabhu joined the channel
#
@Lockedown_ Metadata, Part 2: Microformats https://www.lockedowndesign.com/metadata-microformats/?utm_source=ReviveOldPost&utm_medium=social&utm_campaign=ReviveOldPost (twitter.com/_/status/964904142501285888)
tantek, [jeremycherfas], KartikPrabhu, [kevinmarks] and [cleverdevil] joined the channel
#
KartikPrabhu has there been any consensus on the implied p-name issue?
chrisaldrich and tantek joined the channel
#
Zegnat I think gRegorLove +1’d? I wish we had a quick way to get all the parser implementers to just give a vote.
#
Zegnat But with his +1, it could come to the PHP one, so that’s the biggest (?) one covered.
#
tantek Zegnat is there anything unclear or we could improve about the microformats2-parsing spec change control to make this easier to understand / make happen?
#
Zegnat I don’t think so? I think it is a good thing we (people theorising over the spec) don’t unilaterally change the spec but wait for implementers to comment / express intent. E.g. now there was consensus between aaronpk, you, and me. But none of us actually implement a parser.
#
tantek correct
#
Zegnat Could we have a group on gitHub we can @-mention? Like @parser-implementers? Just something to ping people when the theory people are done debating?
#
tantek usually you want the parser folks involved in the debating, that's part of it
#
Zegnat Yeah, I just don’t actually talk with them often enough, I guess. I couldn’t even tell you who half the implementers are.
#
tantek That's worth knowing though, if you want to seriously help improve the spec, because those are the people you need to convince!
#
Zegnat Maybe I should say that I don’t ever see most of them on IRC, therefor they are outside my bubble, he
#
KartikPrabhu I might already have a working version for mf2py
#
KartikPrabhu now have to install all the libraries so I can test it :P
#
KartikPrabhu <sigh> why is there a Flask requirement for a mf2 parser!
#
KartikPrabhu the frontend should be different from the parser code
vivus and gRegorLove joined the channel
#
KartikPrabhu ok have the stopped implied name working in mf2py. What was the consensus on versioning?
#
tantek you got it working? do you have any test cases for the new behavior vs old?
#
KartikPrabhu I was using the cases reported in https://github.com/microformats/microformats2-parsing/issues/6
#
Loqi [tantek] #6 reduce instances when p-name is implied
#
KartikPrabhu there does not seem to be "expected output" associated to those so if tantek or Zegnat have examples with expected output of new rules that would be great
#
tantek KartikPrabhu: test case meaning something isolated and simplified (perhaps starting from one of the cases reported)
#
tantek that's why it's better to simplify them to a minimal test case that only tests the desired change, rather than a lot more
#
KartikPrabhu yeah. I am not sure what is the expected output though
#
tantek KartikPrabhu: presumably the expected output changes before/after
#
KartikPrabhu yes.
#
tantek so start with minimizing the test case to just what the code change affects, then gather expected output before/after using different parsers (including yours with the patch)
#
KartikPrabhu tantek: if you could add some simple example to that issue with expected before/after outputs that would be great
#
tantek the report
#
tantek KartikPrabhu: that's what I started with asking you for :)
#
KartikPrabhu since I haven't been really following the motivation for this change I am not sure what the new expected output is
#
tantek "do you have any test cases for the new behavior vs old?"
#
tantek then how can you be sure that your change to the code is doing what you think it should be doing?
#
Zegnat <div class="h-entry"><span class="e-content">Hello</span></div> - this should currently give a name and content property, both of "Hello". After the change there should be no name property and only a content property.
#
tantek Zegnat: that's one of three right?
#
Zegnat Is the easiest case I can come up with.
#
tantek https://github.com/microformats/microformats2-parsing/issues/6#issuecomment-366308533
#
Loqi [aaronpk] Nope, wasn't intending to hold things up, just wanted to put that there for the record. I agree with the current proposal of having `p-*` and `e-*` and `h-*` stop the implied name.
#
Zegnat Indeed tantek
[jeremycherfas] joined the channel
#
KartikPrabhu Zegnat: testing now
#
KartikPrabhu Zegnat: could you add those simple examples to the issue with old and new expected outputs?
#
KartikPrabhu I could test all of them pretty quickly before making the change on github
#
Zegnat I could, yes. Uuh. Give me like 10 minutes to construct something for all 3 cases.
#
KartikPrabhu Zegnat: yup no worries
#
KartikPrabhu Zegnat: yup confirmed the example you mentioned works as expected
barpthewire joined the channel
#
Zegnat Almost exactly 10 minutes? :P https://github.com/microformats/microformats2-parsing/issues/6#issuecomment-366473390
#
Loqi [Zegnat] At @kartikprabhu’s request here are some super simplified examples of HTML where unwanted metadata (a `footer` of an `article`) would be swept up in the `name` even though I only want `content` or `like-of` to show. Before and afters have been incl...
#
KartikPrabhu Zegnat: thanks will check those
#
tantek Zegnat++
#
Loqi zegnat has 8 karma in this channel (173 overall)
#
Zegnat Tried to make them somewhat relatable to real-world cases we have seen
#
KartikPrabhu Zegnat++
#
Loqi zegnat has 9 karma in this channel (174 overall)
#
KartikPrabhu my code currently fails only the third one so will fix that
#
KartikPrabhu boom! done passes all three test cases!
#
KartikPrabhu is not sure if he should update version numbers since my repo is not canonical
#
tantek that's something I never understood about git/github
#
tantek once you push to your fork, is there some way of pushing that upstream to the repo you forked from?
#
tantek too confusing
#
KartikPrabhu yes there is. But then tommorris has to merge them
#
tantek and/or move his repo, but I think pending pull requests get moved along with a repo when it gets moved
#
tantek so either way, go ahead with pushing upstream like that
#
tantek and maybe it'll get merged after the repo gets moved - maybe even by you if you want to be a co-owner of the repo at the indieweb org!
#
KartikPrabhu so I guess using semantic versioning this will be bumped to 2.0.0 from 1.0.5 ?
#
KartikPrabhu does not know these rules
#
tantek it's not a feature change
#
tantek I'd guess 2nd digit at most
#
KartikPrabhu ok 1.1.5 it is
#
gRegorLove KartikPrabhu++ for mf2py updates!
#
Loqi kartikprabhu has 4 karma in this channel (166 overall)
#
Zegnat I’d say it is a major release, as everything non major is backwards compatible in semver land, right?
#
KartikPrabhu Zegnat: who knows. don't understand vague rules :P
#
KartikPrabhu this does not change the "API" i.e. the way you interact with the parser
#
gRegorLove Since the "name" property isn't guaranteed in all parsed results now, seems like a breaking change
#
gRegorLove Similar to how adding 'rel-urls' and removing 'alternates' in this php-mf2 commit is a breaking change https://github.com/indieweb/php-mf2/pull/125
#
Loqi [gRegorLove] #125 Add rel-urls object
#
KartikPrabhu hmmm
#
Zegnat Yeah, I am never 100% sure what counts as API and what does not, under semver. Is API just “function Z accepts an input Y” or is it “function Z always returns output X for input Y”.
#
KartikPrabhu since my code is not canonical anyway I am not sure it matters at the moment (?)
#
gRegorLove True, this decision matters once it's merged and a new mf2py release is put out.
#
gRegorLove So not immediately, but "soon" hopefully :)
#
KartikPrabhu does 1.0.5 become 1.1.0 or 1.1.5 ?
#
gRegorLove 1.1.0
#
KartikPrabhu oh ok
#
KartikPrabhu this new math is weird
#
gRegorLove Everything "below" the number you're incrementing gets reset to 0 afaik
#
Zegnat yes
#
Zegnat “Patch version MUST be reset to 0 when minor version is incremented.” - semver.org
#
KartikPrabhu now looks up git commands
#
Zegnat Could always switch to TeX’s versioning, if you want to annoy people.
tantek joined the channel
#
KartikPrabhu no! I stay away from TeX conventions if I can
#
Zegnat (TeX’s version number is \pi. For every release, they add another digit of precision to the end.)
#
KartikPrabhu <sigh>
#
KartikPrabhu ok uploaded mf2py implied name fix to https://github.com/kartikprabhu/mf2py/tree/implied-name-fix
#
KartikPrabhu feel free to test and what not
[tantek], [miklb], KartikPrabhu and [kevinmarks] joined the channel
#
[kevinmarks] I'll have a look. I'd like to redo my connection of the tests in the microformats repository to mf2py
tantek and [cleverdevil] joined the channel