#microformats 2018-03-17
2018-03-17 UTC
# 
[kevinmarks] The problem is usually u- but I haven't seen an actual property collision
[kevinmarks], tantek, KartikPrabhu, nitot, Garbee, cbfishes, [kim_landwehr], Kyle-K and noi278 joined the channel; Kyle-K left the channel
# 
KartikPrabhu experimental mf2py https://github.com/kartikprabhu/mf2py/tree/experimental now uses JSON from https://github.com/kartikprabhu/backcompat-rules to do all backcompat parsing including rels
cbfishes, tantek and [kevinmarks] joined the channel
# [kevinmarks] Nice - that's a simple structure that parallels the mf2 layout
[snarfed], [kevinmarks], gerwitz, [eddie], [cb], [dshanske], [chrisaldrich], [gerwitz], chrisaldrich and [cleverdevil] joined the channel
# 
gRegorLove Issue to discuss allowing numbers in vendor prefixes: https://github.com/microformats/microformats2-parsing/issues/28
# KartikPrabhu thanks [kevinmarks] might use it in v1.1.1
# KartikPrabhu sknebel: alias?
# KartikPrabhu oh ha!
# KartikPrabhu i thought you were talking about the repo
# gRegorLove KartikPrabhu: I think this is an issue in mf2py as well https://github.com/indieweb/php-mf2/issues/155#issuecomment-373957273
# KartikPrabhu yeah just saw your comment
# gRegorLove Using unmung; not sure how up to date tha tis
[snarfed] joined the channel
# KartikPrabhu gRegorLove: you can try my parser https://kartikprabhu.com/connection/mfparser
# KartikPrabhu which is bleeding edge
# KartikPrabhu but it actually is a bug on https://github.com/kartikprabhu/backcompat-rules/blob/master/backcompat-rules/hreview.json
# KartikPrabhu which were mostly imported from old mf2py/php-mf2 rules
# gRegorLove Ah, ok
# KartikPrabhu gRegorLove++ good catch
# gRegorLove So does that mean mf2py is replacing "fn" with "p-item h-item p-name" before parsing?
# KartikPrabhu yes which is also funny
# KartikPrabhu hreview backcompat is complicated by children stuff
# gRegorLove php-mf2 is close to being able to use that json, but it currently uses an additional "context" so it knows how to process some of those properties that get nested
# KartikPrabhu aah interesting
# gRegorLove Failing in this case, though :)
# KartikPrabhu it would be good if somehow those nesting things can be represented in the JSON thing. My attempt was a simply dump of mf2py/php-mf2 transformations along with some additional rules I found missing
# gRegorLove Let me refresh my memory of some of this php-mf2 code. Hopefully it can be simplified to use the backcompat-rules
# KartikPrabhu cool
# KartikPrabhu gRegorLove: if it helps https://github.com/kartikprabhu/mf2py/blob/experimental/mf2py/backcompat.py
# KartikPrabhu no nested children stuff yet
# KartikPrabhu oh! I can do the nesting thing with a h-item JSON dictionary
# KartikPrabhu isn't sure if it makes sense to make backcompat for "item" apply everywhere instead of just inside hreview !?
# KartikPrabhu gRegorLove: is that the context thing you were talking about ^?
# gRegorLove kind of, but this is a special case.
# gRegorLove if there's fn, photo, and url, it's not just a class replacement. They all need to appear under a single p-item h-item.
# gRegorLove the php-mf2 'context' on backcompat comes into play when upgrading say 'item vevent'. It adds 'p-item h-event' and calls the backcompat recursively on the element with 'vevent' as the context, so it knows to only upgrade vevent properties.
# KartikPrabhu hmm yeah hreview backcompat is such a special case
tantek joined the channel
# gRegorLove Yeah, not sure how to do it currently in php-mf2 without adding elements, which sounds messy