#microformats 2015-06-06

2015-06-06 UTC
#
KevinMarks ohm so the test transforming summary ti name was correct?
#
KevinMarks OK, I'll change the parser and revret my yest change
#
KevinMarks s/yest/test/
#
Loqi KevinMarks meant to say: OK, I'll change the parser and revret my test change
#
tantek vevent summary -> h-event p-name
#
tantek pretty sure that's in the spec
#
tantek yes it is http://microformats.org/wiki/h-event#Parser_Compatibility
#
KevinMarks ah, OK
#
KevinMarks so the h-resume issue is that the new rule that the name becomes the value is stomping the old concatenated value
#
KevinMarks and that coudl go weitehr way really
#
tantek probably need to look at some actual classic hResume in the wild to come up with a more specific proposal
#
KevinMarks similarly https://github.com/kevinmarks/tests/blob/master/tests/microformats-v1/hcalendar/combining.json
#
KevinMarks the value of location would now be "Geoloqi" from the name of the h-card
#
KevinMarks is that better?
#
tantek looking
#
tantek wait - are you implementing kylewm's proposal? microformats.org/wiki/microformats2-parsing-brainstorming#Nested_h-.2A_objects.27_.22value.22_property
#
tantek http://microformats.org/wiki/microformats2-parsing-brainstorming#Nested_h-.2A_objects.27_.22value.22_property
#
KevinMarks kyle did implemnt it yes
#
KevinMarks so I got ti when I updated
#
KevinMarks now I'm looking at test regressions
#
tantek well that's my bad with being behind on incorporating into spec then - since we have consensus on that proposal (AFAIK) + one real world implementation
#
KevinMarks also doing leading/trailing space stripping on name now
#
KevinMarks which does help, but i think collapsing space runs within woudl also be good
#
KevinMarks hm, the various ways of hiding properties in geo are going to be more work
#
tantek "think … would also be good" is not the way to reason about it at this point - because it's easy to come up with such arguments either way
#
tantek geo itself is likely unnecessary now that we flattened its properties out
#
KevinMarks abbr, value-title,
#
KevinMarks right, this is backcompat parsing
#
tantek "hiding properties in geo" is not a use-case - it's plumbing, a means to an ends for an actual use-case. so ask yourself, what is that actual use-case, and (re)start from there
#
KevinMarks I'm going through the mf1 test cases
#
tantek at this point, you really should question them unless they cite a real world URL from which they were derived.
#
KevinMarks that may be a useful separate effort
#
tantek questioning is always useful in such ways
#
tantek I'm saying rather than using test case oddness to suggest spec changes, you should question the test case in the first place
#
tantek unless the test case cites a real world example
#
tantek ok I've got microformats2-parsing open in an editable state (as well as -issues and -brainstorming) - going offline for a while to see if I can make all edits for the issues/brainstorming that have consensus and 1+ real world implementation
#
tantek back in a couple hours or so.
#
KevinMarks hm. find bug in testharness that stops when a test file has invalid json
#
KevinMarks suddenly more tests fail
#
KevinMarks on the other had I fixed a lot of tests
tantek joined the channel
#
tantek edited /microformats2-parsing (-120) "incorporate resolved issues: deduping of rels and include alternates in rels" (view diff)
#
tantek edited /microformats2-parsing-issues (+1624) "resolved deduping of rels and include alternates in rels, incorporated into [[microformats2-parsing]], split drop alternates to new issue" (view diff)
#
tantek KevinMarks ^^^
#
KevinMarks aha, so you first to win for type etc as well as text
#
KevinMarks I'd suggest putting the rel-urls clause above the alternate one
#
KevinMarks it's clearer wihtout that if in the middle
#
KevinMarks in practice the rels and rel-urls goes on together
#
KevinMarks https://github.com/kevinmarks/mf2py/blob/master/mf2py/parser.py#L273
#
KevinMarks in particular it easier to keep the rels->urls and urls->rels mapping if you do it one loop like https://github.com/kevinmarks/mf2py/blob/master/mf2py/parser.py#L286
#
KevinMarks !tell tantek read feedback above
#
Loqi Ok, I'll tell him that when I see him next
#
kylewm KevinMarks: that thing with parse errors stopping the test cases is annoying, i fixed a bunch of them in this PR https://github.com/microformats/tests/pull/23
fuzzyhorns and andicascadesf joined the channel
#
KevinMarks___ I think I fixed them all in mine
#
KevinMarks___ Then I merged your unicode stuff and made that a yielding test and now I have 200+ fails
KevinMarks_, KevinMarks__, KevinMarks___ and tantek joined the channel
#
tantek hey Loqi
#
Loqi tantek: KevinMarks left you a message 2 hours, 52 minutes ago: read feedback above
#
tantek KevinMarks yes good suggestion
#
tantek edited /microformats2-parsing (+0) "move now at-risk alternates part of algorithm to end to keep rels and rel-urls creating together per feedback from KevinMarks. change helps clarity without changing functionality" (view diff)
KartikPrabhu, tantek and elf-pavlik joined the channel
#
@SplashCopy Are microformats the most underrated #SEO tool at your disposal? https://www.splashcopywriters.co.uk/blog/microformats-are-they-the-most-underrated-seo-weapon-of-all-time.html (twitter.com/_/status/607088036871041024)
#
KevinMarks you can refactor that algorithm to be clearer
#
KevinMarks do the rel-urls stuff first
#
KevinMarks then iterate the rels adding url to the top-level rels and rel to the rel-urls[url][rels] inside the loop
#
KevinMarks that makes the bidirectionla mapping more explicit
chiui, eschnou, andicascadesf and Left_Turn joined the channel
#
@AllTheTwits is guessing that Jeremy's article is about Microformats :D (twitter.com/_/status/607178898930057216)
elf-pavlik, KartikPrabhu, fuzzyhorns and fuzzyhor_ joined the channel
#
@hmans @hirsebirse Wir haben uns über Microformats2 vs. Microdata gestritten und wurden damit Sujet eines Talks (ich übertreibe.) (twitter.com/_/status/607221739635023872)
KevinMarks_, KartikPrabhu and tantek joined the channel
#
tantek !tell KevinMarks, kylewm ok I think I've incorporated all the resolved microformats2-parsing issues/brainstorms that you guys have implemented in at least one parser. Now need your implementation attempts at items "Status: resolved" here at top of http://microformats.org/wiki/microformats2-parsing-issues
#
Loqi Ok, I'll tell them that when I see them next
gRegorLove joined the channel
#
KevinMarks_ We do parse link in mf2py - currently we are ending up with an empty text in rel-urls. Should that be omitted?
#
Loqi KevinMarks_: tantek left you a message 12 minutes ago: ok I think I've incorporated all the resolved microformats2-parsing issues/brainstorms that you guys have implemented in at least one parser. Now need your implementation attempts at items "Status: resolved" here at top of http://microformats.org/wiki/microformats2-parsing-issues
KevinMarks__ joined the channel
#
KevinMarks eg: http://www.unmung.com/?html=%3Cbase+href%3D%22http%3A%2F%2Fexample.com%22%3E%0D%0A%3Clink+rel%3D%22updates+alternate%22+type%3D%22application%2Fatom%2Bxml%22+href%3D%22updates.atom%22+%2F%3E&pretty=on
KartikPrabhu joined the channel
#
KevinMarks should we omit 'text' for a self-closing link?
Atamido, eschnou, tjgillies_, KevinMarks_ and KevinMarks__ joined the channel
#
KevinMarks grrr something is turning the whole output into ascii instead of uicode
#
KevinMarks s/uicode/unicode/
#
Loqi KevinMarks meant to say: grrr something is turning the whole output into ascii instead of unicode
#
KevinMarks somethign being me passing the json instead of the dict to the cehck code
#
KevinMarks o_O
#
KevinMarks hm, so I may have overdone the 'make evrything unicode' then
#
KevinMarks does from __future__ import unicode_literals make all literal strings unicode?
#
KevinMarks yes it does
#
KevinMarks well that makes a load of changes unnecessary
#
KevinMarks note to self: check tests are failing for the reason you think they are before changing lots of code
#
KevinMarks I did a pull request mainly to get the travis tests run
KevinMarks__, iwaim and rhiaro joined the channel
#
KevinMarks and the travis tests fail because we don't pass all the core microformats tests yet.
#
KevinMarks I'm going to have to do a branch, aren't I
benborges joined the channel
#
kylewm KevinMarks: the way I approached it was to put --exclude in the .travis.yml config
#
Loqi kylewm: tantek left you a message 4 hours, 19 minutes ago: ok I think I've incorporated all the resolved microformats2-parsing issues/brainstorms that you guys have implemented in at least one parser. Now need your implementation attempts at items "Status: resolved" here at top of http://microformats.org/wiki/microformats2-parsing-issues
#
kylewm cause IMO there's not much point in having travis run a bunch of tests we know will fail
#
kylewm https://github.com/kylewm/mf2py/blob/test-suite/.travis.yml
#
kylewm I also put the test harness in its own file: https://github.com/kylewm/mf2py/blob/test-suite/test/test_suite.py
#
KevinMarks I made it call your unicode test thing as a yield so you get a line per failure
#
KevinMarks https://github.com/kevinmarks/mf2py/blob/master/test/test_parser.py#L418
#
KevinMarks and added it into the otehr test harness too
#
KevinMarks ah, I see
#
KevinMarks maybe we should reorganise the mf tests so that the solid ones are separated from the volatile ones
#
KevinMarks could you then exclude the subfolders that are flaky
#
KevinMarks and by merging your test fix I now have conflicts in mine
#
KevinMarks derp
#
KevinMarks OK, if you want to review my changes to https://github.com/microformats/tests/pull/24 have a look
KevinMarks_ and KartikPrabhu joined the channel