#microformats 2018-03-25

2018-03-25 UTC
[cleverdevil], rodney1, [kevinmarks], [miklb], [snarfed], barpthewire, tantek, [eddie], nitot, KevinMarks and webchat254 joined the channel; webchat254 left the channel
#
Zegnat
Only three of php-mf2’s own tests fail when I switch over to my plain text extraction algo!
KevinMarks and Garbee joined the channel
#
aaronpk
tries to catch up on all the php-mf2 progress
KartikPrabhu joined the channel
#
Zegnat
Shouldn’t be too much to catch up on aaronpk. The PRs are pretty self-contained.
#
aaronpk
Zegnat++ for this comment, I would not have thought of that https://github.com/indieweb/php-mf2/pull/164#issuecomment-375959695
#
Loqi
zegnat has 16 karma in this channel (188 overall)
#
KartikPrabhu
aaronpk: Zegnat: to avoid that mf2py makes a copy of the entire mf1 backcompat root and works only on that
#
aaronpk
makes sense
#
KartikPrabhu
I accidently closed that PR
#
KartikPrabhu
how do I undo it
#
KartikPrabhu
oh ok reopen
#
Zegnat
You should have a button to open it as well, if you closed it
#
KartikPrabhu
when did I get so many rights on php-mf2
#
KartikPrabhu
<muhahaha>
#
Loqi
hahahaha
[eddie] joined the channel
#
Zegnat
Has happened to me so often. I really do not understand the Comment & Close butotn.
#
aaronpk
yeah "comment and close" seems like it should not be a button for PRs
#
aaronpk
makes sense for issues
#
KartikPrabhu
lol I have more rights on php-mf2 than on mf2py :P
#
Zegnat
You are added to the mf org, I guess, KartikPrabhu. Unlike me, I couldn’t even at reviewers to my PR :P
#
Zegnat
s/at/add/
#
KartikPrabhu
yeah I am in indieweb/core so I have rights to php-mf2
#
Zegnat
There are a few ways we could get around the contaminating HTML with temp elements thing. We already need special logic for the innerHTML functionality needed for e-* parsing. Could easily be extended to not include certain elements. Like data elements with a class “php-mf2-backcompat-data” or something.
#
Zegnat
Huh. That’s interesting. Did you just merge master back into my e-parsing fix, aaronpk?
#
Zegnat
Wouldn’t rebasing have been neater?
#
aaronpk
I used the "fix conflict" tool in github's UI
#
aaronpk
I had to hand-edit the file to merge the changes from master with yours
#
aaronpk
so it made a new commit
#
aaronpk
the comment "merge master into.." isn't entirely accurate
#
Zegnat
Oh that’s interesting! I haven’t seen that UI yet
#
aaronpk
it's relatively new
#
aaronpk
heh "relatively"
#
Zegnat
I mostly just have private repos for small tidpits of my own code, and would have fixed all conflicts locally for those already
#
Zegnat
Nice stuff though. Solving conflicts may be the most tedious thing with git. (I want to test andto see if it really is better in pijul/darcs.)
#
GWG
If you are going to start merging rel as backcompat, does that include syndication?
#
KartikPrabhu
GWG this is only for rel-tag as described for hentry and hreview backcompat parsing
#
KartikPrabhu
not generic rel values
KevinMarks joined the channel
#
Zegnat
You mean like parsing rel="syndication" as u-syndication on posts, GWG? Do you have examples of such hentry’s?
#
Zegnat
I don’t think that is covered in the backwards compat parsing, but if you have examples of such posts, might be worth adding?
#
KartikPrabhu
by "such posts" of course it means posts without mf2 roots
#
Zegnat
Yep, hentry-no-dash :)
#
GWG
I am wondering about whether it is a good idea. I merge them in my code that consumes php mf2 output
#
KartikPrabhu
GWG: unless there are examples out there it isn't a good idea to add things to backcompat parsing
#
KartikPrabhu
I don't think rel-syndication was even a thing during mf1 days
#
GWG
I am thinking of WordPress sites with old themes that have some mix, but have not looked for an example in the wild
#
GWG
The legitimate problem with WordPress is too many working parts
#
KartikPrabhu
GWG: all backcompat rules are ignored if there is a mf2 root, so unless they are using purely mf1 with rel-syndication
#
GWG
I will start cataloging oddities as per request
#
KartikPrabhu
if you find any examples you can start an issue at https://github.com/microformats/microformats2-parsing
#
Loqi
[microformats] microformats2-parsing: For collecting and handling issues with the microformats2 parsing specification: http://microformats.org/wiki/microformats2-parsing
KevinMarks, tantek and kaushalmod joined the channel
#
gRegorLove
Good catch on the added <data> elements, Zegnat.
#
gRegorLove
I kinda copied mf2py but didn't realize it was working on a copy of the elements.
#
gRegorLove
Adding a special class that's ignored later might be the simplest solution currently, without a big refactoring of the code.
#
gRegorLove
Because the backcompat() function is going through mf1 classes and adding mf2 class names where necessary, then the parser runs on it as if it's authored mf2
#
gRegorLove
This raises a similar issue I hadn't thought of before: http://pin13.net/mf2/?id=20180325183710569
#
KartikPrabhu
mf2py works on the copy only for baccompat
#
gRegorLove
Not sure if there's mf1 like that in the wild, though.
#
gRegorLove
KartikPrabhu, makes sense.
#
KartikPrabhu
oh hmmm i think mf2py will also break on that
#
KartikPrabhu
yup it does
#
gRegorLove
Obviously need some real life examples before we spend time on that. p-name is probably a bad example, but there might be other mf1 properties.
#
KartikPrabhu
right not fixing it now
#
gRegorLove
glances at WordPress
#
KartikPrabhu
mf2py will remove the p-name since it does that to ignore mf2 properties
#
GWG
WordPress is glancing back, gRegorLove
#
gRegorLove
I knew before looking that irc notification was probably GWG since I mentioned the magic word :)
KevinMarks_ joined the channel
#
GWG
It is only one of the magic words
#
Zegnat
That backcompat stuff is hard, which is why I like focussing on some new stuff. Like white spacing ;)
#
Zegnat
has a new “aaronpkExpectations” test
#
gRegorLove
Yeah, it is. I keep thinking we're "almost done" then this stuff comes up. :) It is getting better, though!
#
Zegnat
Hmm, I am breaking 3 php-mf2 tests still with my plain text extraction. So now to figure out if their plain text value is being improved or not by the new algo
webchat140 joined the channel
#
gregorlove
edited /hentry (+1434) "Add Wayback Machine links for dead atomenabled.org references"
(view diff)
#
gRegorLove
Confirmed: I entered "1" for that ReCaptcha^.
#
gRegorLove
Now I'm curious *how long* it's been doing that - if it's just since the shutdown notice or earlier, heh
#
gRegorLove
!tell aaronpk Can you add Zegnat to the indieweb github team so we can assign reviews?
#
Loqi
Ok, I'll tell them that when I see them next
#
Zegnat
I believe it is since the shutdown notice.
#
Zegnat
Btw, why is php-mf2 under indieweb, and mf2py under microformats?
#
gRegorLove
I'm not sure most of us were aware of microformats repo at the time php-mf2 was set up. Think it used to be under barnabywalters, then since the most use was indieweb, it got moved there.
#
gRegorLove
Or maybe not. maybe it was always under indieweb.
#
Zegnat
Should we normalise on one place?
#
gRegorLove
Maybe. I don't really mind either way; not opposed.
#
gRegorLove
Hmm, I might have a simpler solution for the rel-tag added elements. Add it as a direct child of the hentry element. So it shouldn't conflict with the entry-content element.
#
gRegorLove
tries it
#
gRegorLove
Might still cause issues with nested mf1
#
gRegorLove
Appears to work! Just a whitespace issue in the test aaronpk added
#
gRegorLove
Yeah, it works, but still fails if there's an hentry.entry-content nested in an hentry.entry-content.
#
gRegorLove
Odd enough of an edge case that it's ok?
#
Zegnat
entry-content in entry-content is odd for sures, probably safe to ignore for now
#
Zegnat
Ah crap, we missed a thing.
#
Zegnat
goes to do one last PR of the weekend
KevinMarks joined the channel
#
Zegnat
One line hotfix that should probably land before marking 0.4.2: https://github.com/indieweb/php-mf2/pull/166
#
Loqi
[Zegnat] #166 Really fixes #159 - make types unique
tantek, KevinMarks_, [kevinmarks] and [eddie] joined the channel
#
KartikPrabhu
tantek: could you clarify https://github.com/microformats/microformats2-parsing/issues/27 ? Basiclly should ordinal dates be converted to MM-DD in the final step of http://microformats.org/wiki/value-class-pattern#Date_and_time_parsing
#
Loqi
[gRegorLove] #27 vcp: Clarify ordinal dates in parsed result
#
tantek
good q
#
tantek
yeah I think that's only reasonable interpretation of the current text in that last step of VCP date time parsing
#
tantek
and I think I made that choice deliberately to simplify the data for consuming code
#
tantek
rather than all consuming code having to deal with ordinal dates
#
KartikPrabhu
hmmm converting ordinals to MM-DD is tricky due to leap years and all that
#
tantek
what no
#
tantek
goodness sakes it's like a one line simple math expression
#
KartikPrabhu
DDD = 060 is Feb 29 in a leap year but Mar 01 in non-leap years
#
tantek
needs to dig up all his old HyperTalk code and open source it as CASSIS
#
tantek
KartikPrabhu: I'll leave it as an exercise for you to find the simple code to deal with leap years
#
KartikPrabhu
I don't need to python does it
#
KartikPrabhu
in any case, I think would be good to make the spec explicit that this conversion should be done in the last step
#
tantek
why? has it caused any bugs
#
KartikPrabhu
no because it has caused implementor confusion
#
tantek
ah I see in the issue now both gRegorLove and Zegnat were confused :)
#
tantek
but no one has actually implemented it that way right?
#
gRegorLove
No parsers convert ordinal to MM-DD afaik
#
tantek
no parsers accept ordinal AFAIK
#
KartikPrabhu
I just added ordinal dates to mf2py yesterday without conversion
#
tantek
ah so this is about to be a problem
#
gRegorLove
php-mf2 accepts ordinal
#
tantek
really?
#
gRegorLove
I think mf2py does as well
#
KartikPrabhu
mf2py experimental only
#
KartikPrabhu
so I think it would be good to just be explicit about conversion or not
#
KartikPrabhu
I can add the conversion to mf2y too
#
tantek
question is whether released parsers disagree or not
#
gRegorLove
I was incorrect, php-mf2 doesn't do ordinal for VCP
#
KartikPrabhu
i think released parsers both mf2py and php-mf2 don't do ordinals
#
gRegorLove
I think my previous test was only with explicit value in <time datetime=>
#
KartikPrabhu
but to add ordinal parsing it would be good to have some agreement so future releases do agree
#
tantek
right that's what I thought. so I can still fix this as editorial (not breaking any existing parsers)
#
KartikPrabhu
checks other parsers on microformats.io
#
tantek
that shows a failure to parse
#
gRegorLove
"I was incorrect..."
#
KartikPrabhu
tantek: Go parser converts to MM-DD, Ruby does not, php-mf2 and mf2py (released version) do not parse ordinals
#
KartikPrabhu
unsecure connection. me stops
#
gRegorLove
Doesn't get the dt-end though, heh
#
gRegorLove
I know. It's safe though.
#
KartikPrabhu
nope. personal policy ;)
#
gRegorLove
I mean, don't enter your SSN or CC ;)
#
gRegorLove
Someone who's in contact with Glenn should check with him about that.
#
tantek
Ruby parser accepts ordinal though?
#
tantek
and produces it?
#
gRegorLove
Node parser accepts ordinal and produces it.
#
tantek
yikes
#
gRegorLove
I can add the output to the mf2 parsing issue
#
Loqi
[kartikprabhu] Current status of parsers Go: accepts ordinal dates and converts to MM-DD in the final output. Ruby: accepts ordinal dates but does not convert to MM-DD php-mf2: does not accept ordinal dates mf2py: does not accept ordinal dates mf2py (experim...
#
gRegorLove
KartikPrabhu++
#
Loqi
kartikprabhu has 20 karma in this channel (188 overall)
#
KartikPrabhu
tantek: summariesed
#
tantek
edited /value-class-pattern (+37) "assembled dates must be normalized YYYY-MM-DD (because parsed output consuming code likely expects that, either way, simpler for JSON consuming code), proposed since parsers disagree"
(view diff)
#
gRegorLove
+1 for normalized YYYY-MM-DD
#
tantek
commenting now
#
tantek
commented
#
tantek
and put PROPOSED text in spec
#
KartikPrabhu
is that PROPOSED to get feedback from other implementors?
#
tantek
once we've got broader (more people) agreement / consensus on the issue, I'll remove "PROPOSED" from spec accordintly, citing that
#
tantek
we have consensus now but only among three of us (even if two are parser devs)
#
KartikPrabhu
yes fair enough
#
tantek
needs more broader review / agreement
#
tantek
edited /value-class-pattern (+89) "link proposed to issue to comment / provide feedback on #27"
(view diff)
#
tantek
but that should be enough for you to move forward with implementing
#
gRegorLove
Zegnat will be +1 based on previous conversation
#
gRegorLove
I'll file an issue for php-mf2
#
gRegorLove
!tell aaronpk updated https://github.com/indieweb/php-mf2/pull/164, can you run CI again? PHP5.6 didn't get the updated test somehow https://travis-ci.org/indieweb/php-mf2/builds/358164691
#
Loqi
Ok, I'll tell them that when I see them next
#
Loqi
[gRegorLove] #164 rel=tag hentry/hreview backcompat
#
gRegorLove
That issue is the only one open for php-mf2 milestone 0.4.2 currently. I'm +1 for releasing it after that's resolved.
#
sknebel
Zegnat had some fix for a previous PR too
#
gRegorLove
Just added the 0.4.2 milestone to those PRs
#
KartikPrabhu
updated ordinal parsing in mf2py experimental
KevinMarks, [cleverdevil], chrisaldrich and KartikPrabhu joined the channel