#microformats 2024-03-08

2024-03-08 UTC
gRegor and [snarfed] joined the channel
#
gRegor
Cleaning up some php-mf2 PRs, can I get reviews/OK to merge for:
#
Loqi
[preview] [gRegorLove] #250 Fix invalid property class name parsing
#
gRegor
An unfortunate thing I ran into again with Github Actions is that they stop running if you don't have commits for more than 60 days
#
gRegor
Even when they're set to trigger on a new PR. Odd.
#
gRegor
It's set so we can manually trigger it (under Actions), but that only runs it on the main branch, can't select to run it on those PRs
#
[Joe_Crawford]
OOh tests! Nice!
#
gRegor
test all the things!
#
[Joe_Crawford]
:metal:
[jeremycherfas], [Ana_R], sebbu2, [tantek], [qubyte], barnaby, samboyer, [lcs] and to2ds joined the channel
#
sknebel
gRegor: soo... set up a Github action on a schedule to push an unused branch every 30 days?
to2ds, [Paul_Robert_Ll] and [schmarty] joined the channel
#
jkingweb
Talk of tests reminded me I never wrote that README for my unit tests, so I've done so now:
#
jkingweb
The PR should therefore be ready to merge now:
#
Loqi
[preview] [JKingweb] #126 [WIP] Unit tests
to2ds, [Tilley] and jacky joined the channel
#
jacky
so http://php.microformats.io/?id=20240308174917855 confirms this understanding, that property classes are _not_ parsed as if it implied to itself if it's on an element but _would_ do so if it's a child element (but then not for itself but for the owning item of that property)
#
jacky
this distinction was hard for me to grok (and has been probably why the rust parser has been so out of whack)
#
jacky
hah that _might_ have fixed a recursion bug
#
gRegor
Not sure I follow jacky, which implied parsing were you expecting in that example HTML?
[Joao_Paulo_Pes] joined the channel
#
gRegor
Oh, maybe the h-card.name ending up as the implied h-entry.name?
#
jkingweb
I think Jacky is saying that it was unclear that p-, u-, e- and dt- classes are ignored on root h- microformats, that only descendent elements are parsed for properties.
#
gRegor
h-entry doesn't imply a name in that case because it has a p-* property
#
gRegor
Aha, missed the p-content.
#
gRegor
Makes sense
[calumryan] and paulrobertlloyd joined the channel
#
jacky
yeah my fault, I used the word "implied" here differently here
#
jacky
but now that's a potential "lint" I can probably catch as an extension to the parser
#
[tantek]
people mistakenly putting property class names on the same element as a root class name and assuming it belongs to that root is one of the more common errors (it doesn't, such a property is how you embed an entire microformat object *as the value of that property on its parent microformat object*)
#
[tantek]
it's literally how you do things like "this is the location h-card for this h-event" e.g. class="p-location h-card", which btw means the same as class="h-card p-location" since the class attribute is an unordered space delimited set of class names
#
[tantek]
similarly class="p-author h-card" for indicating "this is the author h-card for the parent h-entry"
#
gRegor
Yeah, could be a good thing for something like indiewebify.me to try to catch that
#
[tantek]
^ yes, worth filing a feature enhancement request on indiewebify for that, for both the h-card validator and h-entry validator
#
gRegor
Thanks for the reviews, [Joe_Crawford]!
#
[Joe_Crawford]
Thanks for moving that repo forward!
#
gRegor
Good point on the Portland address. I'm open to whatever
#
[tantek]
I figured Powells would be fine because they're effectively an institution that's going to be around, and seem positive to send them some additional discoverability
#
[tantek]
Keeps the theme of Portland in the example
#
[tantek]
Also if nitpicking about examples is annoying, feel free to say so 🙂
#
gRegor
I don't mind :)
#
gRegor
I'll fix the indieweb.org links in this one, then a separate issue to fix them and https://example.com in the rest of the tests
to2ds joined the channel
#
gRegor
The extra commit should get tests to run too, so that's good
paulrobertlloyd, jacky and barnaby joined the channel
#
[tantek]
apologies gRegor, reload 😂
#
gRegor
Need to file an issue on microformats/tests too, these tests were copied from https://github.com/microformats/tests/blob/master/tests/microformats-v1/hcalendar/attendees.html
#
gRegor
So no apologies needed, good to catch these in the main microformats tests!
#
gRegor
Updated, confirmed the actual event was 2012-10-10
#
[tantek]
DRY++ MOISTT--
#
Loqi
MOISTT has -1 karma over the last year
#
Loqi
DRY has 1 karma over the last year
#
[tantek]
mistakes occur in saying things twice
#
[tantek]
(yes just made up that backronym)
#
[tantek]
also ironic that the *visible* "Nov" was the incorrect month, and the ISO8601 date in the datetime attribute was correct!
#
[tantek]
(normally the visible content tends to be more accurate than the code / invisible content)
#
gRegor
yeah I was surprised, initially just updated the datetime then did some searching to find the event
#
[tantek]
feel free to ping me on updates to the tests or other places that have that datetime inconsistency bug
#
[tantek]
updates as in PRs etc.
[Joe_Crawford], [Fran], plantroon and [campegg] joined the channel
jacky joined the channel
#
[tantek]
nice! gRegor++
#
Loqi
gRegor has 6 karma in this channel over the last year (103 in all channels)