#wordpress 2018-08-22

2018-08-22 UTC
[iambismark] and Typhon15 joined the channel
#
Typhon15
| ____
tsp11 joined the channel
#
tsp11
| 0 .-' '-. 0 1 0 0 0
#
tsp11
| 11 0 / _ _\ 1 0 1 1 10
[matpacker] joined the channel
#
[matpacker]
Good morning everyone. Been away on holidays the last 2 weeks, have I missed anything big?
#
GWG
Such as?
#
[matpacker]
New theme support, new plugins, etc.
#
GWG
I'm still trying to get Micropub out the door
#
GWG
Looking for beta testers
#
[matpacker]
What's required? I should be able to help out.
#
GWG
[matpacker]: Install the Github version of Micropub. Try it
#
miklb_
master branch
#
GWG
miklb_: I created a pull to try to solve your problem
#
miklb_
solve which problem? I have so many
#
GWG
The one on the line you commented out
#
miklb_
all of which are my own making
#
Loqi
[snarfed] wordpress-micropub: A Micropub server plugin for WordPress
#
miklb_
oh I'll check it in sec
#
GWG
[matpacker]: Yes
#
miklb_
set up the mastodon autoposter see how that works out.
#
miklb_
I need to put some helper scripts on my server to make it easier to checkout pull-requests in repos that aren't mine
[chrisaldrich], [cleverdevil], tantek__, ben_thatmustbeme, [eddie] and [matpacker] joined the channel
#
[matpacker]
GWG: is there much of a difference between the Git hosted Micropub and the Wordpress one? The WP one was working for me, but the Git one doesn't.
#
Ruxton
[matpacker]: they're seriously diverged atm, it looks like GWG is splitting the code out in master at present, current version on WP is about here https://github.com/snarfed/wordpress-micropub/tree/979b19d8e8cab88c3469b306a4113d1dff10de6d
#
Ruxton
note the difference being it is all in one php file at that commit, after it starts to be broken out into includes dir, etc.
#
Ruxton
also https://github.com/snarfed/wordpress-micropub/releases usually shows the point at which releases to WP are made
#
[matpacker]
Ruxton: ahh, ok... Will have to compile the error messages
globbot and puff10 joined the channel
#
puff10
__
[kevinmarks] and Alex`27 joined the channel
#
Alex`27
.. _ .. ""..
Davesdowntime and host joined the channel
#
host
____
jeremych_ and By joined the channel
#
By
.-' '-.
#
By
/ _ _\
#
By
/':. (o) /__)
ben_thatmustbeme, petermolnar and [kevinmarks] joined the channel
#
GWG
Ruxton, didn't just split the functionality, did a lot of refactoring
jeremych_, deltab29 and [jgmac1106] joined the channel
#
[jgmac1106]
does anyone else micro.blog from the 2106 theme? trying to find out if my problem with the rss feeds is just me or universal
#
[jgmac1106]
@gwg Chris said adding a follow post kind is easy as it is just commented out in the plugin. I am in the editor but do not see it.
#
[jgmac1106]
@sknebel I didn't think I had to since I saw the follow.svg in the plugin, though I didn't find the template file in the view folder
#
[jgmac1106]
just remember Dave and Chris saying the follow post kind is there it is just commented out. or something
[keithjgrant] and [grantcodes] joined the channel
#
miklb_
[jgmac1106] I'm using 2016 with the mf2 plugin, not the fork. What issues are you having?
#
[jgmac1106]
the rss feeds
#
miklb_
what about them?
#
[jgmac1106]
it is on my radar to work on this week since i get a few days off before summer
#
[jgmac1106]
the html is being stripped on some, I want to change the bookmark and quote post kind feeds to have the title
#
[jgmac1106]
but mainly for now if I add <a href="https://micro.blog/miklb">@miklb</a> the feed strips away the link
#
[jgmac1106]
here is my quote feed, missing link and title and <blockquote> http://jgregorymcverry.com/kind/quote/feed/
#
[jgmac1106]
I know what needs to be fixed and how to do it, just needed some capacity
#
miklb_
I gotcha.
#
miklb_
In related news, I'm really feeling good about 2016 is performing with just the mf2 plugin and the 2 small additions I've made
#
miklb_
Still soliciting feedback on the h-feed parsing, but I have mastodon auto-poster and have checked my posts in Together via the json feed and the atom feed in Evergreen, everything looks good and the syndication posts look good. I haven't checked micro.blog yet
#
miklb_
this is what a micropub reply to twitter URL in Quill looks like syndicating to mastodon https://mastodon.social/@miklb/100593897909903585
#
Loqi
[miklb] In reply to shep on Twitter. Yeah, now that I’ve set up indiepaper.io for read later, I’ll probably do the same. I was using open in safari as a read later hack on iOS ... https://miklb.com/blog/2018/08/22/4187/
#
miklb_
unless I'm missing something, that seems like a pretty good experience using the IW plugins out of the box with a core theme.
#
sknebel
https://miklb.com/blog/2018/08/22/4187/ has a second name property with the value "Previous Post", but apparently the consumers ignore that in favor of the first, empty one?
#
Loqi
[Michael Bishop] In reply to shep on Twitter. Yeah, now that I’ve set up indiepaper.io for read later, I’ll probably do the same. I was using open in safari as a read later hack on iOS
#
sknebel
(PTD says something else, but I'm not sure how many things follow it strictly)
#
aaronpk
i don't think i take into account multiple values of properties that i expect to have only one value
#
sknebel
I suspect most don't
#
[jgmac1106]
[miklb] and this is no post kinds correct?
#
miklb_
p-name had been mentioned before and multiple values, but so far no one has mentioned any broken parsing so I was just working my way down the list til I had specific issues.
#
miklb_
no, this is with the post-kinds plugin
#
[jgmac1106]
but that is good news. My other question would be on threading twitter conversations, in my experience only bridgy publish handles this well
#
miklb_
I'm using bridgy publish too. Just switched to the core 2016 theme, added the uf2 plugin & have inserted 2 new elements
#
miklb_
I made no other changes to my existing IW plugin setup besides that
#
[jgmac1106]
even better.....can you try an rsvp? bookmark? and quote? and share those results. Those can be finnicky in Bridgy publish
#
[jgmac1106]
ooohhh okay, I thought Quill did it, should have realized that was bridgy
#
[jgmac1106]
but I am using the forked version no mf2 plugin
#
[jgmac1106]
I will publish the same and we can compare
#
miklb_
https://miklb.com/blog/2018/08/20/4167/ is a bookmark butI didn't syndicate it because even with my custom mf2 theme I had issues with sending it to Twitter
#
Loqi
[Michael Bishop] A little appropos seeing this quote shared by Matt today after I wrote this last night “WordPress was orginally about making it easier to put words and pictures on the Internet. Gutenberg and 5.0 is about ushering in a new interface to do that. We ...
#
[jgmac1106]
yeah I listened to the recode/decode podcast last night, heard same line
#
miklb_
so the bookmark syndication isn't indicative of a mf2 theme issue I believe than how the template is constructed in the post-kinds view template as it relates to bridgy
#
[jgmac1106]
Good, sounds like a documented issue then
#
[jgmac1106]
I guess no point in me publishing a comparison
#
miklb_
no, by all means compare. I just noted that I had a custom mf2 theme and still had issues getting a bookmark to syndicate nicely through bridgy
#
[jgmac1106]
though the quote kind with a tweet url works perfectly
#
miklb_
may be a case of using a custom excerpt for those
#
miklb_
yes, I wrote the quote template for bridgy-publish specifically to work with Twitter quotes and the expected bridgy markup
#
[jgmac1106]
bridgy can do no wrong for me, only tool I ever found that allows me to keep thread alive on Twitter and webmentions
#
Loqi
miklb has 22 karma in this channel over the last year (38 in all channels)
#
[jgmac1106]
and using bridgy and twitter is the only way I can send a webmention, everything else has to be manual
#
[jgmac1106]
I have to learn more about what Alan is soing with Splot's: http://splot.ca/splots/splotbox/
#
sknebel
miklb++ for comprehensive testing and doing something about the results
#
Loqi
miklb has 23 karma in this channel over the last year (39 in all channels)
#
miklb_
I just think GWG and pfefferle underestimated how good the work they done is. 2 minor shims I've added?
#
miklb_
they've done
#
miklb_
is woefully under caffeinated
#
miklb_
that should have also read I wrote the quote template for post-kinds to specifically work with Twitter quotes as expected by bridgy
#
sknebel
reading back through the initial discussion of problems: plugins adding share buttons etc to the post body and that being included in the mf2 contents and thus syndicated etc
#
sknebel
I believe we've seen issues with simple location doing that too
#
sknebel
although maybe the hook hierachy solves that?
#
sknebel
If I read it right the mf2 plugin tries to be the first to add anything, so it only wraps the actual post content in "e-content"?
#
miklb_
correct
#
sknebel
e.g. jmac1106 has often "see also on:" or something in his tweets because of something like that?
#
sknebel
and the thing with the recent comments sidebar, let me see if I can find an example again
#
miklb_
those seems unavoidable even in a custom theme the way WP is constructed and the limited hooks to inject content
#
miklb_
the sharing buttons etc
#
[jgmac1106]
yeah that just happened to me, I was recording tutorials on making custom menus
#
[jgmac1106]
I put the "social links" menu there and it remained even after removing
#
miklb_
I'm using the syndication links plugin that adds the icons to where the content is syndicated and it doesn't interfere with the content
#
Loqi
[Ton Zijlstra] Wrapping My Head Around Webmentions Pt 2
#
miklb_
what am I look for?
#
sknebel
the "url" property
#
sknebel
that's the leaking recent comments widget thing
#
[jgmac1106]
I can easily recreate the recent comments sidebar issue,
#
[jgmac1106]
if you enable it the hcard from the comment bleeds into the post
#
[jgmac1106]
but I need to get to work, grant due tomorrow, need to finish entire package
#
miklb_
that is SemPress theme from pfefferle so that goes beyond adding mf2 via a plugin. I'm not suggesting we don't fix it, but goes to show that even in a custom mf2 theme there are still problems that can arise
#
sknebel
would be interestng to me if the mf2 plugin causes the same problem
#
sknebel
since it also does mark up comments, or if it somehow is better about that
#
miklb_
I can add the recent comments widget and start testing
#
miklb_
mind you that widget doesn't handle empty post titles
#
sknebel
okay, but it doesn't happen
#
sknebel
good to know
#
miklb_
I'm seeing the name issue on the notes without titles from that previous post link now.
Peng_19 joined the channel
#
Peng_19
| ____
#
miklb_
so what reader is available to view my site by it's h-feed
#
aaronpk
you should be able to use aperture+monocle
#
GWG
There is no template for follow yet. Next Post Kinds release I will write one. It would use the generic template
#
GWG
I try to build things as modular as possible
#
GWG
I want to be like Ikea
[grantcodes] joined the channel
#
miklb_
aaronpk thanks, I'll set up Monocle
#
aaronpk
you should be able to install the aperture plugin in wordpress and then log in to monocle.p3k.io
#
miklb_
yes, I set up the aperture plugin already for using with Indigenous and Together. ah, OK.
#
miklb_
ok, the p-name thing needs to be fixed for sure. Awesome
#
miklb_
awesome meaning it's a tangible issue I see "broken"
#
sknebel
where does it cause issues?
#
miklb_
either an issue in my mf2 or an issue with Monocle, and I am always going to lean toward it being mine before aaronpk https://cdn.miklb.com/Main_2018-08-22_10-47-22.png
#
aaronpk
whoa what the heck
#
miklb_
that doesn't look like that in Evergreen or Together
#
miklb_
same note I mean in the other 2 readers
#
aaronpk
or in the browser
#
miklb_
heh, maybe so
#
miklb_
that is Safari Tech Preview 12.1 so…
#
aaronpk
i'm confused, because xray doesn't show that either https://xray.p3k.app/parse?url=https%3A%2F%2Fmiklb.com%2F&pretty=true
#
sknebel
that's in the entry-title
#
sknebel
but that never should be parsed
#
sknebel
and in the rel=bookmark, but same thing
#
Zegnat
I think you subscribed to your JSON feed
#
Zegnat
That’s the item title there: view-source:https://miklb.com/feed/json/
#
aaronpk
that makes a lot more sense
#
Zegnat
I am not sure if JSON Feed titles are allowed to contain HTML. So not sure what the proper handling is. It could be other readers do not know either. Some might be stripping the HTML, some might be rendering it (rendering an empty string), and others (apparently like Monocle) endcode the < and > for display
#
aaronpk
title in jsonfeed must be plain text https://jsonfeed.org/version/1
#
Zegnat
Yep, I was just about to write the same :)
#
Zegnat
So Monocle is displaying the right thing.
#
miklb_
so should I open an issue or check for an existing one in the wp json feed plugin repo?
#
miklb_
or still an mf2 problem to try and solve?
#
aaronpk
this is definitely a problem with the jsonfeed plugin
#
aaronpk
or wait
#
aaronpk
actually
#
aaronpk
does wordpress allow html in the title?
#
aaronpk
if yes, then it's a jsonfeed plugin problem. if no, it's a mf2 plugin problem.
#
aaronpk
htmlentity decode... doesn't that do things like turn &lt; back into <
#
miklb_
WP allows HTML in titles, at least something like <strong> and <em> I don't know if it has a white list or not
#
sknebel
then that seems like the jsonfeed plugin should strip that
#
sknebel
although if it's possible to modify the mf2 plugin to not modify the title in this case, that'd be good too
drdanick26 joined the channel
#
drdanick26
________________ _ . ____ *
#
miklb_
I actually can modify that function in the mf2 plugin.
#
miklb_
I should be able to also check if $title is empty and do something else there if need be
#
sknebel
with "in this case" I meant "when called for feed generation" , it seems like that's not exposed
#
sknebel
but since the in-build feeds demonstrate that html has to be stripped, that can probably be expected from plugins that want plaintext titles
#
miklb_
aaronpk so can I remove the json feed in aperture and use h-feed instead?
#
aaronpk
yeah i thought we were trying to troubleshoot the html feed
#
miklb_
I'm troubleshooting all the things, but thinking out loud
#
miklb_
oops, I left the old entries, should have removed them too.
tantek__, gRegorLove and jackjamieson joined the channel
#
miklb_
a note looks fine
jackjamieson joined the channel
#
miklb_
should an issue in the json feed plugin be opened then? I was sure the determination
#
Zegnat
miklb_, probably. The JSON Feed plugin could be expected to do the same as WordPress’ own RSS feed code, which is stripping this HTML from titles. That would solve the issue. Just reference the lines in WP core that sknebel found.
#
sknebel
yeah, I'd say they probably can reuse the title cleaning functions from the wordpress feeds
[jgmac1106], [keithjgrant] and miklb joined the channel
#
tantek__
miklb, I'm catching up on your analysis
#
tantek__
what were the two minor edits?
#
Loqi
[Michael Bishop] But as it stands, with only two minor edits to the mf2 plugin & suite of Indieweb WP plugins I already had installed, I’m using default twentysixteen WordPress theme. I’m syndicating to micro.blog, Twitter (via brid.gy) and Mastodon (via auto-pos...
#
tantek__
miklb++ for testing all this
#
Loqi
miklb has 24 karma in this channel over the last year (40 in all channels)
#
miklb
I added a second <time> with dt-published and added that hidden data element for the u-url to the mf2 plugin
#
GWG
I asked Manton to give me commit privileges to jsonfeed as he's busy, but I think he forgot
#
miklb
I'm seeing little things that are still wonky using the h-feed for microsub, but nothing that so far screams broken.
#
miklb
we surfaced a jsonfeed issue which was good
#
miklb
it's still pretty vanilla, but I didn't remove anything I had installed.
#
GWG
I should chase Manton again
#
miklb
I just replied to my own note that had been syndicated to Twitter and it threaded perfectly as Twitter reply. Those are the kinds of experiences I think people are looking for, and to see it work through a new lens is exciting.
#
GWG
Anything I need to do?
#
miklb
If we want to merge those 2 new things I believe there is a PR open. But we might be able to tweak a couple more things.
#
GWG
Which things?
#
miklb
but not really. The nature of WordPress and plugin/theme conflicts is as old as the platform. That's not going to magically change adding IW tools to it. But having a solid foundation to work from is really impressive.
#
miklb
the time and data element that I added.
#
miklb
[12:21:49] <+miklb> I added a second <time> with dt-published and added that hidden data element for the u-url to the mf2 plugin
#
miklb
that's all I've changed in all of this testing.
#
GWG
Didn't I submit the time one?
#
miklb
if you did, I didn't see it in master before I started doing this.
#
miklb
I didn't think to look for open PRs
#
Loqi
[dshanske] #23 Spec says date in RFC3339 format but nothing about GMT
#
miklb
no, mf2 plugin
#
miklb
I'm adding a second <time class="dt-published">
#
miklb
as well as a data element with the u-url permalink as the value
#
miklb
goes in the h-entry before e-content
#
miklb
so with those 2 things added to the mf2 plugin twentysixteen is working without any edits thus far in real world testing
#
GWG
I didn't do that PR
#
miklb
I'm confused. which PR?
#
GWG
Anything with date and uf2
#
miklb
I didn't mean to suggest you did. I did it. It's in the GitHub repo now.
#
miklb
looks like though its a mixed PR now but they are both in there https://github.com/indieweb/wordpress-uf2/pull/39
#
Loqi
[miklb] #39 feature: add u-url to h-entry
#
miklb
I'm editing live on my server as well as locally and just wanted to make sure I pushed the changes to the repo, so I need to squash or clean up the PR. But the point is, that's all I've modified
#
miklb
we have been telling people they need to install this special theme, or this forked theme for a GitHub repo. Now I think we could comfortably say, "Install WordPress, activate all of these IW plugins and follow their instructions (brid.gy for example) and use the stock 2016 theme."
#
miklb
or we are really close to being able to say that.
#
miklb
that seems pretty cool to me. You should be proud.
#
jackjamieson
Agreed, miklb! The WordPress system is making bounds these days
#
Loqi
GWG has 74 karma in this channel over the last year (152 in all channels)
#
miklb
and generally speaking, most 2016 child themes should just work™ Anything else wouldn't really be a child theme anymore
#
miklb
if I'm being overly optimistic let me know though.
#
GWG
I like optimism
#
GWG
By the way, re your IndieAuth issue, did you notice the new developer documentation?
#
miklb
I haven't looked at that much as I've been focusing on the front end stuff. I have a busy rest of the day but this evening I will look at that PR.
#
miklb
micropub PR
#
GWG
It basically removes the conditional that you somehow tripped. I didn't need to, but no one is using it
[grantcodes] joined the channel
#
miklb
I read through the PR discussion a bit, I'll give it more attention when I test. Thanks.
#
GWG
I am working on Client Information Discovery for IndieAuth. Basic implementation
#
miklb
I have no idea what that is
#
GWG
So, getting the application name and icon
#
GWG
Instead of just saying quill.p3k.io it would say Quill
[cleverdevil] joined the channel
#
GWG
So, simple enough feature
#
GWG
I also have to do something with Refbacks
#
GWG
Miklb, did you ever try Refbacks? They are only useful for non Webmention supporting sites
#
miklb
no haven't
[kevinmarks] and dougbeal|mb1 joined the channel
#
tantek__
miklb re: extra <time> for dt-published, if we added backcompat for 'entry-date' to h-entry, then all mf2 parsers would pick that up automatically
#
tantek__
so that would eliminate one change, and make things work "out of the box" for more users
#
miklb
well I'm not sure how wide spread entry-date is used
#
tantek__
that's what I was hoping you might be able to determine - you saw it by default in 2016 right?
#
miklb
that's what is in this core theme yes
#
tantek__
it doesn't have to be universal, if it is widespread *enough* then it's worth back-compat parsing for
#
tantek__
also entry-date as a logical addition to entry-title, entry-summary, entry-content makes sense
#
miklb
and that is a theme specific function generating that part of the html, thus the shoe-horn of another time element
#
miklb
but sure, if that can be eliminated from the markup that would certainly help (the extra time element)
#
tantek__
and then the data element was for URL / permalink right?
#
miklb
u-url
#
tantek__
then I'm confused because that should already be picked up as part of backcompat handling of rel=bookmark which I see in 2016
#
miklb
got me. That was the first thing aaronpk pointed out as seeming broken
#
tantek__
this may be a bug with the specific mf2 parser
#
tantek__
rather than any fault of WP / 2016
#
aaronpk
i hadn't considered that tbh haha
#
aaronpk
i thought the php parser had solid backcompat
#
tantek__
likely depends on if we have an hentry rel=bookmark backcompat test case
#
miklb
but it also has h-entry if that makes a difference
[pfefferle] joined the channel
#
[pfefferle]
Is the backcompat working if the container is h-entry?
#
[pfefferle]
If you do not activate the mf2 plugin, it works...
#
tantek__
wait I thought 2016 didn't have h-entry
#
aaronpk
ohhhh yea that's what's going on
#
miklb
it doesn't. The mf2 plugin adds it
#
aaronpk
we're not talking about a pure backcompat case
#
miklb
right
#
aaronpk
we're talking about making the mf2 plugin work
#
tantek__
ok I thought we were close to getting it all working with a pure backcompat case *without* mf2 plugin
#
tantek__
that is, getting 2016 working *as is*
#
tantek__
without a plugin
#
[pfefferle]
That is what I tried to say the last two days
#
miklb
hmm. I think would miss out on the other stuff by not having mf2
#
[pfefferle]
2016 is doing fine without the mf2 plugin
#
tantek__
miklb, were you able to document precisely what was missing *before* adding the mf2 plugin?
#
miklb
doing fine is again a very subjective term in this scenario. If you don't have mf2 then you loose a lot of the the other features of the IW plugins
#
miklb
so 2 shims seems like a small price to get mf2 working. IMVHO
#
tantek__
miklb, point is that you should be seeing enough from backcompat to get those IW plugins to work
#
tantek__
miklb - I'm hoping to approach 0 shims
#
miklb
it's all a bloody shim if you really look at it.
#
tantek__
not quite
#
tantek__
when we add backcompat rules, it's not a shim
#
miklb
if there is no h-feed or h-entry, will it pick up the mf2 in the entry?
#
tantek__
there are two separate sets of questions
snarfed joined the channel
#
tantek__
1 what core hentry/h-entry properties are missing in 2016? (dt-published would be handled by issue 7 proposal)
#
tantek__
2 what additional mf2-only properties would we need to enable additional use-cases / plugin features?
#
tantek__
let's focus on (1) first
#
miklb
I didn't test that because I was going under the assumption that we 100% needed mf2 to get all of the Indieweb goodness.
#
tantek__
I believe solving even just (1) is highly valuable
#
miklb
I'm happy to remove the mf2 plugin and keep testing if there is a consensus upstream that can be solved across the board.
#
tantek__
hence why I keep asking for thumbs-up/down and comments on https://github.com/microformats/h-entry/issues/7
#
Loqi
[tantek] #7 Can we improve h-entry backcompat to handle common WordPress hentry streams?
#
[pfefferle]
As I said
#
[pfefferle]
I thought this was already defined in backcombat, because php-mf2 supports it already
#
tantek__
pfefferle - did you read issue 7?
#
miklb
well, again, this is the first I'm hearing of any option of mixing mf1 & mf2.
#
tantek__
pfefferle - then did you have an opinion on the proposals there?
#
miklb
so this is a whole new conversation
#
tantek__
miklb, "mixing mf1 & mf2" is about (2) above
#
tantek__
let's focus on (1) first
#
tantek__
not a new conversation, trying to refocus the original question
#
[pfefferle]
Yes, got my +1
#
miklb
no, because the other WordPress Indieweb plugins are going to be outputting mf2, so it can't really be done separately in a vacuum
#
tantek__
thanks [pfefferle]
#
tantek__
miklb, that's an oversimplification
#
miklb
I think I'm missing something
#
tantek__
that's the point of (1). by doing *just* those fixes we we enable *some* use-cases
#
tantek__
I would like to know how far we can go
#
tantek__
even if we can't go to 100%
#
miklb
OK. So what do I need to do? Turn off all of the IW plugins to see how it parses totally vanilla WP?
#
tantek__
because some of those use-cases are *very* useful, e.g. if we get hfeed support to be as good or better than RSS, then readers can universally prefer h-feed
#
tantek__
(that's the point aaronpk keeps complaining about)
#
miklb
well there is still the idea on the table of the mf2feed
#
[pfefferle]
for readers we currently have no problems using the php-mf2 parser
#
tantek__
sure, ok to pursue mf2feed in parallel
#
tantek__
[pfefferle]: that's incorrect on at least a couple of points, per aaronpk feedback
#
tantek__
it's exactly why I put the two proposals into issue 7
#
aaronpk
latest example of a post showing up wrong https://i.imgur.com/5b5PBM5.png
#
aaronpk
that's a webmention even, not hfeed
#
tantek__
aaronpk - not helping
#
[pfefferle]
No, aaronpk only tested the version WITH the mf2 plugin
#
[pfefferle]
But that kills the backcompat
#
tantek__
so either folks want to focus on specific cases at a time, or we're going to get nowhere
#
miklb
aaronpk do you have the link to the original post?
#
tantek__
because everyone is confusing all the different cases
#
[pfefferle]
Aaronpk can you please test this site with your reader https://twentysixteendemo.wordpress.com
#
tantek__
right, now aaronpk you changed the subject to permalinks when we were talking about feeds
#
[pfefferle]
This is a vanilla 2016 theme
#
tantek__
so I give up for now
#
miklb
no, I want to understand tantek__
#
miklb
I'm game to test anything live. I'm just confused what I need to turn off.
#
tantek__
miklb, yes you do, but aaronpk is not helping with "that's a webmention even, not hfeed"
#
aaronpk
[pfefferle]: all but the last two look great
#
tantek__
point of (1) and (2) above is to start with focusing on feed testing
#
miklb
define which feed :-)
#
tantek__
miklb, homepage hfeed
#
aaronpk
[pfefferle]: oh the featured image is missing
#
aaronpk
i thought i support featured images in monocle
#
miklb
does it need the class to pick it up?
#
aaronpk
also what is that subtitle text? that isn't coming through, not sure whose fault that is
#
aaronpk
"image post format" is missing the image
#
[pfefferle]
This is the vanilla version using mf1
#
aaronpk
"gallery post format" works great
#
[pfefferle]
Without any u-features
#
[pfefferle]
U-featured
#
aaronpk
good set of test content!
#
tantek__
if we can get at least some consensus that the two proposals in https://github.com/microformats/h-entry/issues/7 are worth trying implementing in a parser, we can iterate and see what else we can add
#
Loqi
[tantek] #7 Can we improve h-entry backcompat to handle common WordPress hentry streams?
#
aaronpk
ah [pfefferle] apparently i forgot to display the summary in monocle. i see those are coming through right in xray
#
tantek__
feel free to thumbs-up the whole thing, or even just indicate that you only like .entry-date or only like rel=author
#
tantek__
I'd like to collect folks's opinions either way
#
[pfefferle]
Is it on purpose that you show the url instead of the author name?
#
tantek__
given that parsers already support rel=bookmark inside an hentry for backcompat, I presume supporting rel=author inside an hentry for backcompat is doable
#
[pfefferle]
Sure, php-mf2 seems to also do that right now
#
aaronpk
[pfefferle]: was that a question for me?
#
tantek__
[pfefferle]: "url instead of the author name" is likely due to lack of rel=author in the backcompat
#
tantek__
hence if folks want to fix that, please say so on the issue
#
[pfefferle]
No, the parser gets the name
#
[pfefferle]
Aaronpk yes
#
aaronpk
i show the url if there is no author name
#
aaronpk
which screenshot are you referring to? the last one I posted has the author name
#
[pfefferle]
But the X-ray output shows the name
#
miklb
I understand what you are attempting to do tantek__ but I guess I have no skin in the game as to making WP mf1 backcompatible in parsing. The user experience of what I outlined in that note requires mf2. I don't see how the 2 are supposed to cross.
#
tantek__
[pfefferle]: pretty sure X-ray does more than just mf2 parsing & backcompat
#
[pfefferle]
Might be, but the plain php lib also supports it
#
[pfefferle]
That’s what I am trying to say since yesterday ;)
#
tantek__
miklb, at least *some* of that ux is possible with just WP mf1 more backcompat, and that's still helpful, andthat's my point
#
[pfefferle]
I have to work on my English skills :(
#
Loqi
aww, cheer up
#
tantek__
miklb, for the additional UX that you think requires more explicit mf2, I'd like to understand what those use-cases are specifically
#
tantek__
so that we might have a chance for looking for the information in existing markup
#
tantek__
(similar to the two proposals in issue 7)
#
miklb
are you saying bridgy syndication and webmention replies won't need mf2?
#
tantek__
this is why I'm not willing to give up yet
[jgmac1106] joined the channel
#
tantek__
miklb, very possibly, yes
#
[jgmac1106]
[miklb] as well as documenting or linking to the two things you changed can you also describe the testing methodology used so folks like me can copy what you did, was there a specific parser, feed reader, should we drop links into any repos for others to review, etc,
#
[jgmac1106]
[pfefferle] or get more people who speak your tongue in the community to balance it out
#
[pfefferle]
miklb yes! Only if you need the fancy reply context stuff of the post-kinds plugin
#
miklb
not really any methodology as there isn't a true pass/fail
#
tantek__
miklb, also, I'd like to solve the hfeed questions first (readers, etc.), and *then* solve the hentry permalinks second (bridgy syndication and webmention replies)
#
tantek__
they are two different areas
#
miklb
post-kinds is still a recommended plugin in the IW plugin.
#
tantek__
the only sane way to do this kind of testing is to explicitly breakdown the separate cases *separately*, and get full understanding of each before moving onto the next
#
miklb
but I love that we've gone from patching the mf2 plugin to making it obsolete by better back compat
#
miklb
tantek__ agreed. Like I said, until about an hour ago, I did not understand that you were proposing not requiring mf2 in WP themes. I didn't fully understand how that worked. I thought you were still looking at ways to use the mixed formats.
#
[pfefferle]
I am frustrated!
#
tantek__
miklb, got it. appreciate the patience as we understand each other better :)
#
tantek__
miklb, to be clear, I think we *can* solve hfeeds *without* any additional mf2 in themes
#
aaronpk
i would be very happy if mf1 backcompat in default wordpress themes worked better, since more sites will "just work" when i try to follow them even if they don't have the mf2 plugin installed
#
tantek__
^^^ this
#
miklb
there are no mf2 in themes unless it's one of the handful of indie themes.
#
tantek__
this is why it is worth solving the hfeed problem first
#
[pfefferle]
What problem?
#
tantek__
[pfefferle]: what aaronpk has kept saying about following / reading WP blogs
#
tantek__
NOT all the webmention commentary. ignore that for now
#
tantek__
ignore anything about hentry permalinks until hfeeds "just work"
#
[pfefferle]
But this is mainly a problem of us recommending the mf2 plugin
#
tantek__
jumping to a solution is never "mainly" about a problem
#
tantek__
just pretend the mf2 plugin didn't exist.
#
[pfefferle]
The mf1 backcompat works well since I started my work on the semantic-linkbacks plugin
#
[pfefferle]
Not perfect but well enough
#
tantek__
I think its existence is being a distraction from this attempt at regathering this data
#
[pfefferle]
It breaks backcompat and does not add all needed data
#
tantek__
[pfefferle]: at this point we need to use specifics instead. what doesn't work? what bar are you meeting that makes you say "well enough"? what features are still missing?
#
aaronpk
i really need to make a feed previewer thing for monocle so you can see what i see
#
[pfefferle]
If all themes are using the mf1 implementation of 2016 and all parser are working like the php one, I see no issues
#
tantek__
aaronpk, can you:
#
Loqi
[tantek] #7 Can we improve h-entry backcompat to handle common WordPress hentry streams?
#
tantek__
2 file another issue on what you see is missing when you try to follow / subscribe a "normal" modern WordPress blog
snarfed left the channel
#
miklb
this is what I've been trying to say as well " at this point we need to use specifics instead. what doesn't work? what bar are you meeting that makes you say "well enough"? what features are still missing?"
#
tantek__
by doing 2, you (aaronpk) can answer [pfefferle]'s question above "What problem[s]?"
#
miklb
I've been approaching that from a user perspective in a live environment with the current available tools to explore how the content is propagating and appearing in various readers.
#
tantek__
like actually file the issue with specifics, feel free to add the screenshot(s) from your reader(s) there, but list the information you think is missing
#
tantek__
because sorry it is not obvious from screenshots
#
[pfefferle]
If the theme uses proper mf1 the php parser has no problems... if the theme uses no semantics we sure have a problem
#
tantek__
list the specific properties you see as missing
#
tantek__
[pfefferle]: "has no problems" is false, hence I filed issue 7
#
[pfefferle]
That’s why I always reference the php parser 😉
#
miklb
what exactly does the php parser do to say "this works"
#
[pfefferle]
It already supports rel-bookmark and rel-author
#
tantek__
[pfefferle]: AFAIK even using the php parser, on a plain WP 2016 install, published and author are still missing
#
[pfefferle]
And the date parsing
#
tantek__
wait since when?
#
[pfefferle]
For quite some time
#
[pfefferle]
That’s what I am trying to say since 2 days
#
tantek__
again, specifics please. as of March it did not.
#
tantek__
and since no one commented on issue 7, I presumed it was all still true
#
aaronpk
I don't see entry-date in the parsed result https://pin13.net/mf2/?id=20180822181556466
#
[pfefferle]
Tantek how have you tested it?
#
tantek__
on plenty of plain WP 2016 blogs
#
tantek__
back then
[chrisaldrich] joined the channel
#
tantek__
[pfefferle]: aaronpk's example proves my point. if the php parser supported entry-date, then there should be a "published" property in the output of the PHP parser
#
[pfefferle]
I am not sure if backcompat supports the time element
#
[pfefferle]
Please try the vanilla 2016 theme with pin13 Aaronpk
#
aaronpk
is that the test site you gave me?
#
aaronpk
i do see the dates there, but that one also uses the <time> element
#
aaronpk
oh it's using "published"
#
[pfefferle]
As the mf1 spec says
#
[pfefferle]
Tantek the backcompat worked since the first version of my Webmention plugin
#
tantek__
I need to see if I can dig-up which common themes I found at the time had .entry-date but not .published
#
tantek__
aaronpk, can you open a new issue just for capturing examples of WP blogs you've tried to follow that had bad results?
#
tantek__
(new issue on h-entry)
#
tantek__
bad or insufficient
#
[pfefferle]
...and do not use the mf2-plugin!
#
tantek__
well, if you can tell
#
tantek__
but yeah, those would be worth their own issue
#
aaronpk
i need to make a preview tool before i can really do that
#
miklb
don't use mf2. You must use an mf2 theme. wtf
#
aaronpk
no that was to me. he was saying make a list of WP blogs that have bad results that do not also use the mf2 theme.
#
aaronpk
s/theme/plugin
#
miklb
no [pfefferle] is still "–and do not use the mf2-plugin!"
#
aaronpk
right that was a qualifying sentence to tantek
#
miklb
and I'm all like this morning, cool, this stuff is all working like advertised 😂
#
aaronpk
combine tantek's sentence and that one
#
aaronpk
"for capturing examples of WP blogs you've tried to follow that had bad results ...and do not use the mf2-plugin!"
#
aaronpk
it was not an imperative
#
miklb
ok, I just know [pfefferle] has been very down on that plugin from the start.
#
miklb
wonder if twentysixteen still takes PRs
#
tantek__
miklb, I for one am not "down on that plugin", I want to see how much we can raise the bar without it first.
#
miklb
right on
Guest45420 joined the channel
#
Guest45420
___
[kevinmarks] joined the channel
#
miklb
I can turn it off on my site right now and leave the other IW plugins as is to see what happens.
#
miklb
if that would help anyone for testing backcompat
#
tantek__
ideally it would be interesting how much we could get working without any plugins, but I understand wanting to keep your site working :)
#
miklb
as long as the content is being published on my site and being stored in my database that I backup, I'm not worried about anything else. If anyone is reading my feed that isn't a developer I'd be surprised.
[schmarty] joined the channel
#
miklb
if it looks wonky on Twitter to get to the bottom of things, that's what I'm all about right now.
#
miklb
I still have indiewp.com on that same server that I can put some test content on and compare with/without mf2 as well
#
sknebel
indiemap had the following sites that might benefit from entry-date parsing: https://gist.github.com/sknebel/f7bdf1d58116a61419add4af97be5438
#
miklb
interesting
#
miklb
I'm also wondering if those 2 shims are all that's needed how easy would it be to just send a PR to that theme to add the 2 classes
#
miklb
but I need to run to the market. tantek__ let me know if turning off or setting up another demo site that's live would help.
#
[pfefferle]
Tantek sorry, not everything was accurate what I wrote... the php parser does not support entry-date and rel-author but the default themes are using proper mf1 for date and author/hcard so that is why it is no problem... sorry, I am on the phone and have not verified the source...
#
[pfefferle]
So I will have to do some research to check if entry-title and rel-author really helps
#
[pfefferle]
I mean entry-date
#
sknebel
maybe checking if the templates used on the blogs on that list are popular is a starting point
#
[pfefferle]
Sknebel, which of them also have the published class??
#
sknebel
I tried to exclude those
#
sknebel
so hopefully none, but can't promise it.
#
[pfefferle]
Only entry-date without published?
#
sknebel
(+ of course sites might have upgraded since the crawl, I removed one such case I recognized)
#
tantek__
good list sknebel. I wonder if we can detect any patterns, like which themes they are using
#
[pfefferle]
Hmmm, these themes are completely broken
#
[pfefferle]
They use h-as-note instead of hentry and have no content
#
[pfefferle]
So I think they are broken mf2 themes
#
tantek__
oh dear, I thought we contained h-as-* to just some of the old plugins that GWG was working on
#
tantek__
notices http://www.manton.org/ in that gist
#
Loqi
Manton Reece
#
GWG
It wasn't me. It was pfefferle
#
GWG
I copied his h-as
#
GWG
Also, what's this about Post Kinds?
#
GWG
On another break
#
[pfefferle]
tantek__ the theme adds h-as-note and that breaks backcompat
#
tantek__
yes we dprecated h-as-* a while ago
#
[pfefferle]
seems to be added by a plugin
#
tantek__
ugh which plugin?
#
tantek__
I thought we fixed all those
#
GWG
Me too
[tantek] joined the channel
#
miklb
maybe it just wasn't updated yet
#
miklb
or someone copied to their theme a custom post-kind template with the old markup
tantek__, jackjamieson, [kiai], miklb_, miklb__, [schmarty], gRegorLove and tantek joined the channel