#microformats 2018-07-24

2018-07-24 UTC
#
snarfed ok, on to the next surprise in recent mf2py changes (eg in 1.1.1 vs 1.0.4):
#
snarfed when it parses this: <div class="h-entry">foo <a href="http://bar">bar</a></div>
#
snarfed it returns properties.url = ["http://bar"]
#
snarfed 1.0.4 didn't
#
snarfed reading http://microformats.org/wiki/microformats2-parsing#parsing_for_implied_properties , evidently that's correct, but it's surprising
#
Loqi [Tantek Çelik] microformats2 parsing specification
#
KartikPrabhu yes, it is correct according to spec. That was a mistake in 1.04
#
snarfed so weird. seems like it would result in a ton of false positives. for any post that doesn't explicitly include u-url, return any arbitrary link inside it?!
#
KartikPrabhu above my pay grade ;)
#
KartikPrabhu if you get this issue on parsing and it has consensus we can change it in 1.1.2
#
snarfed ok. happy to escalate to anyone else here :P
#
KartikPrabhu tantek: ^^
#
[kevinmarks] It's the specific use case for a minimal h-card or h-entry
#
snarfed ok. i get that. just seems dangerous and unreliable in practice.
#
snarfed for h-entries at least
#
snarfed (support for minimalism totally makes sense for h-card. less so for h-entry, at least to me.)
#
[kevinmarks] http://microformats.org/wiki/microformats2-implied-properties
[eddie] joined the channel
#
[kevinmarks] For h-entry it makes sense for an h-feed like page
#
snarfed eh. very minor benefit in syntactic sugar from eliding the u-url class. major potential for false positives, eg when you change your markup and the permalink isn't the first link in the entry.
#
tantek hmm yes that does sound bad
tantek__ joined the channel
#
[kevinmarks] If you have cases in the wild that show this, list them
#
snarfed fair point
#
snarfed i'll refrain from submitting unit tests :P
#
snarfed even without known examples, i'm afraid to deploy this to bridgy, since if someone hits it, they'll start posseing posts that point to arbitrary other urls as the permalink, instead of their own original post
#
snarfed bad failure mode
#
[kevinmarks] Well, part of cleaning up the mf tests repo should be getting rid of arbitrary cases in favour of real ones
#
[kevinmarks] Hm. There are classic feed cases that look like that though - where the permalink is elsewhere
#
KartikPrabhu aah, snarfed good point. there the u-url would be missing from the original post but some other post is linked too before?
#
snarfed sure
#
snarfed ohhh i missed that this only happens when it's the *only* link in the h-entry. ok, i follow the intuition better now. that makes false positives seem much less likely.
#
KartikPrabhu yeah was thinking the same
#
[kevinmarks] http://www.unmung.com/feed?feed=https%3A%2F%2Fdaringfireball.net%2Ffeeds%2Fmain
#
[kevinmarks] (I generate more verbose markup but most of those posts have an external permalink in the u-url, with the local link on the star)
#
snarfed yup, those permalinks are at the end, but due to the "only link" rule, u-url wouldn't be implied, so no false positive
#
snarfed ah and they have explicit u-url too :P heh nm
#
gRegorLove There is a recent example of a falsely-implied u-url in h-entry: https://github.com/microformats/microformats2-parsing/issues/36
#
Loqi [sknebel] #36 Falsely implied u-url
#
tantek I thought I already captured somewhere wanting to do the parallel rule for u-* as we have for implied p-name
#
gRegorLove It was mentioned in the implied p-name discussion, but not explicitly part of that proposed+implemented resolution
#
tantek gosh this is an old discussion: http://microformats.org/wiki/microformats2-parsing-issues#implied_properties_when_an_explicit_class_is_provided
#
tantek yup we already had consensus on this for implied URL back at IWS 2016!
#
tantek "+1 consensus in the room at IWS 2016 (willnorris, gRegor, kylewm, tantek): do this for specifically for implied URL
#
tantek "
#
tantek where "this" in context is: "I think it makes sense for u-url and can be easily added to the spec as e.g. ".h-x>a[href]:only-of-type:not[.h-*,.u-*]". Kylewm 00:46, 1 December 2015 (UTC) "
#
tantek not sure why I didn't make the edit to the spec at the time, probably just fell through the cracks :(
#
gRegorLove Cool. Might have been one of those items that missed getting the spec update. I recall another 2016-ish item like that
#
tantek this is why the spec still says: "Note: This specification is only marked as a "Draft Specification" because of pending edits from resolved issues before 2016-06-20."
#
tantek alright I'm going to make the spec change per the 2016 consensus resolution
#
tantek then if that needs further iteration we can iterate in 36
#
gRegorLove sounds good
#
gRegorLove (this was the other 2016 one that missed spec update: http://microformats.org/wiki/index.php?title=microformats2-parsing&diff=prev&oldid=66660)
#
Loqi [Tantek Çelik] microformats2 parsing specification
#
snarfed tantek++
#
Loqi tantek has 20 karma in this channel (463 overall)
#
snarfed gRegorLove++
#
Loqi gregorlove has 38 karma in this channel (263 overall)
#
gRegorLove I'll look through /microformats2-parsing-issues for any other 2016 items that might not have made it in
#
tantek edited /microformats2-parsing (+70) "restrict implied u-url per http://microformats.org/wiki/microformats2-parsing-issues#implied_properties_when_an_explicit_class_is_provided "consensus in the room at IWS 2016"" (view diff)
#
tantek do we have a separate issue for implied u-photo?
#
gregorlove edited /microformats2-parsing-issues (+226) "/* exclude style elements before parsing */ link to revisions where spec was updated with consensus" (view diff)
#
tantek uh oh
#
gRegorLove tantek: yeah, https://github.com/microformats/microformats2-parsing/issues/21
#
Loqi [aaronpk] #21 reduce instances where photo is implied
#
tantek adding
#
tantek edited /microformats2-parsing-issues (+250) "/* implied properties when an explicit class is provided */ confirm edit, noting remaining u-photo issue" (view diff)
#
@Lucid00 ↩️ @jimpick @ThisIsMissEm @BeakerBrowser I think they got the idea from this. The spec is identical. http://microformats.org/wiki/rel-payment (twitter.com/_/status/1021564891960946688)
#
gRegorLove Another one that didn't make it: http://microformats.org/wiki/microformats2-parsing-issues#u-_parsing_iframe_src
#
gRegorLove will edit, unless you're already in there, tantek
#
tantek go for it
#
gregorlove edited /microformats2-parsing (+32) "/* parsing a u- property */ add `iframe.u-x[src]` per http://microformats.org/wiki/microformats2-parsing-issues#u-_parsing_iframe_src "+1 from the room at IWS 2016"" (view diff)
#
gregorlove edited /microformats2-parsing-issues (+143) "/* u- parsing iframe src */ link spec revision" (view diff)
#
gRegorLove new issues for php-mf2 and mf2py as a result :)
#
tantek nice!
#
tantek gRegorLove++ thanks much!
#
Loqi gregorlove has 39 karma in this channel (264 overall)
#
gRegorLove I *think* that's everything from IWS 2016 updated in mf2-parsing now
#
tantek oh? guess I'll have to review all the issues!
#
tantek that's not too bad then. I thought I had a lot more work to do
#
gRegorLove I didn't double check the ones that explicitly say "spec updated"
#
tantek probably worth a review
#
gregorlove edited /microformats2-parsing-issues (+81) "/* img fallback in p- */ link to spec update revision" (view diff)
#
gregorlove edited /microformats2-parsing-issues (+91) "/* consistent implied name url from grandchildren of root */ link to spec update revision" (view diff)
#
gregorlove edited /microformats2-parsing-issues (+91) "/* need more not h-* to avoid child root implying properties */ link to spec update revision" (view diff)
#
gRegorLove tantek, not sure what the change was on this one, but I think it's in the spec now? http://microformats.org/wiki/microformats2-parsing-issues#implied_date_for_dt_properties_both_mf2_and_backcompat
#
gRegorLove Not sure if there was a resolution here: http://microformats.org/wiki/microformats2-parsing-issues#Noscript_skip.2Fparse
#
KartikPrabhu has to go through spec again!
#
KartikPrabhu is there a recent changes page?
#
KartikPrabhu this would be easier if every spec change came with a test case somewhere
[miklb] joined the channel
#
gRegorLove KartikPrabhu, I linked to the revision in the github issue
#
gRegorLove Oh, need to file one for implied u-*. We have an example of that.
#
gRegorLove Filed https://github.com/indieweb/php-mf2/issues/182 and https://github.com/indieweb/php-mf2/issues/183 and same for mf2py
#
Loqi [gRegorLove] #182 Fix parsing for iframe.u-*[src]
#
gRegorLove Those are the only two microformats-parsing updates from today
tantek__, KartikPrabhu, [snarfed], adactio, [pfefferle], [jgmac1106], snarfed, [schmarty], TallTed, taragon-pdx and [eddie] joined the channel
#
sknebel aaronpk: the wiki wants a gentle shove please
[cleverdevil] joined the channel
#
aaronpk It looks like it up to me?
#
sknebel it's now, yes ;)
#
aaronpk I didn't do anything !
snarfed, [generativist], tantek__, [ryan339] and KartikPrabhu joined the channel; snarfed left the channel
#
KartikPrabhu alright what is the wiki mirror address
#
sknebel https://microformats.org.mirrors.sknebel.net but it's missing the latest changes :/
#
KartikPrabhu do it now when the wiki is on
#
KartikPrabhu aaronpk: "Fatal error: Out of memory (allocated 2097152) (tried to allocate 7680 bytes) in /var/www/html/wiki/includes/GlobalFunctions.php on line 1944"
#
KartikPrabhu on the wiki
#
sknebel I've got a current copy of the parsing spec ready if you need it?
#
KartikPrabhu yeah I want to see the implied url changes
#
sknebel https://www.svenknebel.de/temp/2018-07-24-pspec.html
#
Loqi [Tantek Çelik] microformats2 parsing specification
#
KartikPrabhu thanks! will this be up for a little while?
#
sknebel sure
#
KartikPrabhu sknebel: added some comments to your PR https://github.com/microformats/mf2py/pull/119/files
#
KartikPrabhu sknebel: ok the major objection I had is fine. Could you quickly add the comment before implied url and your PR is good :)
[joewillmott] joined the channel
#
@chrismessina #meta /cc @dens (at @Foursquare HQ in New York, NY w/ @microformats) https://www.swarmapp.com/c/hU5JmMXgfJz https://t.co/tQlBtB4sh4 (twitter.com/_/status/1021847463672602624)
#
sknebel KartikPrabhu: done
#
KartikPrabhu yup waiting for CI checks!
#
aaronpk Looks like Apache needs some tuning https://i.imgur.com/KXZL2mC.png
[ryan339], jgmac1106 and [eddie] joined the channel
#
jgmac1106 hello all only used slack as I could never get #indieweb-microformats to work thx [KarthikPrabhu]
#
jgmac1106 more importantly thx [skenbel] for the mirror site
#
@dens ↩️ @chrismessina @Foursquare @microformats Sorry I missed you! Here on Cape Cod for Crowley Family Vacation™️ (twitter.com/_/status/1021856520659787776)
[kevinmarks], [generativist], tantek__, [jgmac1106], [cleverdevil], [eddie], hadleybeeman, dbaron, [chrisaldrich], strugee, dmcweeney and KartikPrabhu joined the channel