#wordpress 2018-08-20

2018-08-20 UTC
tantek, ski_, miklb and bradenslen joined the channel
#
bradenslen miklb thanks. I might hold off experimenting till after Gutenberg drops.
#
miklb I am able to authenticate and post using micropub 1.4.3 and master of indieauth, but get `Micropub Error: 401 unauthorized - missing access token` using master from Quill.
#
miklb with Quill
#
miklb that's with everything that's been merged into master for both today
suim4 joined the channel
#
tantek__ miklb re: https://chat.indieweb.org/wordpress/2018-08-19/1534713478952300 I am interested in the specifics of your results, in particular towards improving https://github.com/microformats/h-entry/issues/7 by default
#
Loqi [miklb] I'm going to switch my live theme to a core theme and start looking for what doesn't work as it does with my current custom theme.
#
miklb now that I can at least post from Quill on my phone, I will be attempting more content creation. I just updated 2016 on my install and am making the switch shortly.
#
miklb I would like to rephrase "I can at least" to I've found a working combination due to all of my testing.
tantek joined the channel
#
miklb generic 2016 with nothing different about my site except added the mf2 plugin is now live at https://miklb.com
jamesbond and LambdaComplex21 joined the channel
#
aaronpk Wow that theme is missing a lot of mf2
#
aaronpk https://i.imgur.com/avI5gRa.jpg
#
aaronpk Granted I am not fetching post permalinks yet, but this illustrates the problem well
#
miklb that's all I've been trying to get to the bottom of. Couldn't find another way.
#
miklb then again indiewebify.me thinks its ok https://indiewebify.me/validate-h-entry/?url=https%3A%2F%2Fmiklb.com%2Fblog%2F2018%2F08%2F19%2Flive-testing-the-state-of-affairs-in-wordpress-the-indieweb%2F
#
aaronpk The reason those don’t show up well in my reader is because the feed view is missing author information as well as missing the link to the post permalink
#
aaronpk right the permalinks have the info but when the posts are on the home page list view they do not
#
miklb So I'll see if I can find a way to resolve that without editing a theme file.
#
aaronpk I’m subscribed to your home page, and only fetching the info from there. Not fetching post permalinks
#
aaronpk you should be able to see those results in the pin13 parser
#
miklb but that falls into is it broken or is it a basic experience?
#
aaronpk its debatable
#
aaronpk because you could argue that my feed reader should be fetching each post permalink
#
miklb sure. as I said earlier, valid mf2 or basic mf2 doesn't mean it meets user expecations
#
aaronpk on the other hand, rss/atom/jsonfeed does not require fetching permalinks
#
Loqi +1
#
aaronpk So this experience is solidly worse than if I were to subscribe to your RSS feed
#
aaronpk which is why I end up subscribing to the Wordpress RSS feeds when I see that I’m trying to subscribe to a Wordpress blog
#
miklb I understand and don't disagree. Let me see if there's a way to add u-url without editing a theme file.
#
miklb it gets rel=bookmark but no class on the URL
#
miklb and that is done at the theme level
#
miklb so the question is can I insert another u-url into the h-entry that parsers can pick up. Right? Not necessarily alter the one that is there.
#
miklb not another, a u-url with another link
Guest92014 and tantek__ joined the channel
#
tantek__ miklb, aaronpk this is exactly why I want to analyze the markup of default core themes to see if there's more h-feed and/or h-entry backcompat that we can add
#
tantek__ so there's more "by default" that is parsed
#
tantek__ miklb++ for doing this experiment
#
Loqi miklb has 19 karma in this channel over the last year (35 in all channels)
#
tantek__ aaronpk: this is exactly the point of https://github.com/microformats/h-entry/issues/7 (though maybe we need a separate issue for h-feed backcompat improvements)
#
Loqi [tantek] #7 Can we improve h-entry backcompat to handle common WordPress hentry streams?
#
GWG :I'm talking to someone about IndieAuth as an auth provider for WordPress
#
aaronpk oh?
#
GWG He doesn't understand how it solves the registration issue.
#
GWG But he did tell me a better way to do scopes.
#
tantek__ aaronpk I'm looking at miklb's home page h-feed and noticing that his blog posts do have rel=bookmark on them for the permalink URL, which should work in backcompat parsing
#
aaronpk tantek__: but backcompat parsing is disabled if mf2 is found yea?
#
aaronpk GWG: the point is that registration is unnecessary when we use the whole DNS as the registration system
#
GWG That was my answer
#
tantek__ however I noticed that there is explicit h-feed and h-entry classes, so none of the backcompat parsing is happening, and yes, by design
#
tantek__ the point is that if you know enough to put in explicit mf2 classes, then backcompat gets out of your way and assumes you will do all the markup work, instead of something half & half which is less reliable
#
aaronpk GWG: the other trick is that client secrets are useless for mobile apps and javascript apps so there's no benefit to having a secret in the first place
#
tantek__ miklb are you trying to add minimum mf2 to the theme, or trying to figure out how we can better parse the default theme?
#
tantek__ just curious which questions you're trying to answer
#
aaronpk at which point all of the other registration info is public anyway, so might as well advertise it at a URL
#
tantek__ right
#
GWG He went to lunch. We'll see if he responds
#
aaronpk i think i need to write a blog post about this with a controversial title
#
tantek__ aaronpk y tho
#
aaronpk to get peoples attention
#
tantek__ wait, which "this"?
#
aaronpk oauth client secrets are useless
#
tantek__ ohhhh yeah that
#
tantek__ sounds applicable to -dev in general
#
miklb tantek__: I am trying to find out what is "broken" using the existing IW plugins with a non-mf2 theme
#
tantek__ ohhh ok. so that's *with the mf2 plugin running*
#
miklb and see if I can add the fixes into the mf2 plugin, or, know with certainty what is "broken"
#
miklb yes. the defaul 2016 WP theme, with the mf2 plugin and all of the other IW plugins I was already running.
#
tantek__ goodness sakes look at all those body:not(.search-results) .entry-summary selectors in the homepage inline CSS
#
miklb that might be google_pagespeed doing that. I added that until I do proper criticalcss
#
miklb turned on the flag for pagespeed to inline the CSS for me in the cached html
#
tantek__ miklb looks like the mf2 plugin is failing to add "class='u-url'" to elements with rel="bookmark" inside the entry
#
miklb yes, we discussed that. that is being inserted with the theme markup. I want to look to see if I can add another link to the permalink that has u-url. A hidden link.
#
miklb can the u-url be in the e-content/entry-content?
#
miklb entry-content it looks like I can insert another link to via the plugin
#
tantek__ yes it may
#
tantek__ though it will then be propagated inside the value of the e-content since all the HTML inside is preserved
#
tantek__ should be fine, as long as its an empty link. though you may want to use an empty data element instead to be more a11y friendly
#
tantek__ e.g. <data class='u-url' value='https://example.com/permalink'></data>
Ks0stm11 joined the channel
#
miklb I was curious about that ,and data element seems like a good solution. It would be in the entry-content which has e-content inside
#
tantek__ right, data element there would be less intrusive than an empty a element
#
tantek__ miklb is it too hard to re-use the existing a element with rel=bookmark?
#
miklb that would require editing the theme file, which then defeats the mf2 plugin
#
tantek__ does the mf2 plugin run before or after the theme "runs"?
#
miklb before, more or less.
#
tantek__ so it doesn't get a chance to do "fix up" after the theme then?
#
miklb not that I'm familiar with. Unless there is a hook/filter the theme is the last line in output
#
tantek__ yes that makes it much harder
#
miklb ok, so there is a data element with u-url now inside the h-entry.entry-content div
#
miklb but for some reason it is getting wrapped `<p></p>`
#
tantek__ because WP likes to wrap everything in content in <p></p> 😂
#
miklb yes wp_autop has been the great equalizer for all devs at some point.
#
miklb but that seems to look OK in pin13
#
miklb what's next™
#
tantek__ looks like we could improve authorship a little
#
tantek__ maybe
#
tantek__ the author h-card is currently "just" a child of the h-entry
#
tantek__ whereas ideally it would be a p-author / u-author
#
miklb looking at that now
#
tantek__ miklb do you now what is generating <span class="author vcard"> ?
#
tantek__ we'd need to make that <span class="p-author vcard">
#
miklb that's what I'm looking for now.
#
tantek__ at least that's the minimum change, and then still depend on backcompat support for hCard/vcard
#
miklb that is coming from a theme function to output the entrymeta :-(
#
miklb so it's an easy fix to modify/override in a child theme, but nothing that I can see to do in a plugin
#
tantek__ so either we replicate the entire author structure in the plugin, invisibly (yuck), or we try another direction, like fixing up the backcompat of the hentry and hfeed
#
tantek__ so that "author vcard" automatically gets fully picked up as part of backcompat
#
miklb keep in mind this is just one core theme, so no guarantees that all themes use "author vcard"
#
miklb if that was being generated by core, we could expect it most everywhere
#
tantek__ got it
#
miklb but this is exactly the kind of thing I was looking for.
#
miklb there isn't a way to inherit another h-card is there?
[eddie] joined the channel
#
miklb like if I add the h-card widget with a fully marked up mf2 h-card would the parser be able to use any of it in the feed?
#
tantek__ yes I think it could, especially if it is inside the h-feed as a p-author
#
miklb OK, I added the h-card to the page
#
miklb and I need to leave it at that for the day.
#
tantek__ miklb looks like that worked to put an "author" on the h-feed with your h-card
#
tantek__ so now the question is whether an h-feed consumer will pick up the author from the feed
#
tantek__ I'm guessing it/they will
#
miklb 🤞
[eddie] and [pfefferle] joined the channel
#
[pfefferle] What about the p-name? Do you have a solution for that?
#
[pfefferle] An other point h-Feed parsing is only important for feed readers, with the Webmention plugin, we only parse single posts...
Dworf, [metbril] and [pfefferle] joined the channel
#
[pfefferle] sorry, I let you try it yourself! 😞
[kevinmarks] joined the channel
#
GWG Good day
[kevinmarks] and [jgmac1106] joined the channel
#
[jgmac1106] he everyone. Nice efforts on the testing overnight
[jgarber] joined the channel
#
stephenpieper.net edited /Wordpress_IndieAuth_Plugin (+274) (view diff)
[stefp] joined the channel
#
[stefp] Hi. Could someone look at the change I just made to the Wordpress IndieAuth plugin page on the wiki?
#
[stefp] Is there enough info there?
#
miklb !tell [pfefferle] what/where do mean p-name?
#
Loqi Ok, I'll tell them that when I see them next
#
Zegnat [stefp], that is really interesting! Thanks
#
Zegnat That’s definitely something others can work from. Thank you for checking on the exact rule that needed whitelisting!
#
miklb https://indieweb.org/wiki/index.php?diff=51173&oldid=51123&rcid=51048 *
#
miklb oops. still getting hang of using irssi
#
sknebel hm, seems like the provider of those rules has no clear way for non-customers to report false positives
#
sknebel or I can't find it
#
sknebel ( https://wiki.atomicorp.com/wiki/index.php/Reporting_False_Positives )
#
[stefp] I was in contact with the tech person at my host
#
[stefp] he asked me to try the plugin again and he checked the firewall at the same time
#
[stefp] so he could check what was happening
#
[stefp] and it was that rule causing the problem
#
Zegnat It is an anti remote file injection security, so it blocks when external URLs are used for specific arguments. IndieAuth does this all the time because you identify yourself with a URL, the client that tries to log you in identifies itself with a URL, the client provides a URL to be redirected to after login...
#
Zegnat And as sknebel said, we’d need the IndieAuth usecase whitelisted by Atomicorp.
#
[stefp] Thanks. I was trying to find out what rule 340162 was exactly.
#
Zegnat It is https://wiki.atomicorp.com/wiki/index.php/WAF_340162
#
Zegnat I’ll add that to the wiki :)
#
[stefp] Thank you 🙂
#
miklb So pfefferle seemed to mock fixing h-feed parsing but when I made the switch to 2016 & mf2 plugin, the issue mentioned were about the h-feed. So we really can't even as a community decide what is/isn't broken/good enough. That's difficult to sell.
#
miklb but if anyone besides aaronpk does h-feed I'd love some feedback on mine now that I've added u-url and an h-card on the h-feed
[manton] joined the channel
#
sknebel your h-feed has a url of "https://www.brid.gy/publish/twitter"
[aaronpk] joined the channel
#
miklb I need a new irssi theme. Some of the text doesn't contrast well with my shell environment in iTerm2
#
miklb sknebel: where are you seeing this?
#
sknebel when parsing your homepage: http://pin13.net/mf2/?url=https%3A%2F%2Fmiklb.com%2F
#
sknebel let me check the source
#
sknebel there's a <a href="https://www.brid.gy/publish/twitter"></a> at the bottom, seems that triggers implied url parsing
#
sknebel (line 734)
#
miklb gotcha. That's coming from one of th IW plugins so I should be able to solve that.
#
sknebel your posts have two name properties, one from the headline and one from the "Leave a comment on <title>" section
#
sknebel (and posts without title shouldn't have a name at all, or a name the same as the content
#
vanderven.se martijn edited /Wordpress_IndieAuth_Plugin (+1236) "/* Troubleshooting */ Documenting the ModSecurity issue, huge thanks to [[User:Stephenpieper.net]]!" (view diff)
#
Zegnat [stefp], thanks for the follow up, I have updated the wiki with links to the rule and some more descriptions.
[tantek] joined the channel
#
Zegnat If your hosting company is a customer of Atomicorp, maybe you can get them to report the false positive? Possibly have them mention that not just this plugin but possibly all IndieAuth (https://indieauth.net/) implementations will be hit by this.
#
Loqi IndieAuth https://indieauth.net/assets/web-signin-splash.jpg
#
aaronpk Use the w3.org url when sending them info, it looks more legit :-)
#
Zegnat (https://www.w3.org/TR/indieauth/)
#
aaronpk I'm guessing they already have a whitelist for redirect_uri since that didn't trigger the error in my testing and it's part of OAuth 2.0 so it's used everywhere
#
aaronpk they just need to add client_id and "me" to the whitelist as well
#
Zegnat Could be. I am not sure if the rules are freely available from somewhere so we can inspect exactly what they are matching
#
aaronpk Probably not, but that is what you can tell them they need to whitelist, and it is documented in the spec
[eddie] and [pfefferle] joined the channel
#
[pfefferle] miklb p-name == the title of the post
#
miklb ok. but saying "what are you going to do about it" didn't seem like a helpful response. Just feels like more negativity
#
[pfefferle] thats why I apologized
#
[pfefferle] miklb the problem of the hidden data in the e-content is that it is part of the HTML and will be included in the target site… so it might cause issues parsing the target
#
miklb it's not in e-content, it's in entry-content. I don't know if that makes a difference
#
[pfefferle] but I was really curious
#
[tantek] Good catch sknebel. There should never be a need to put a Bridgy publish url in an stream / h-feed
#
[pfefferle] miklb do we talk about the backwards compatibility? will it be parsed as e-content?
#
miklb that brigdy url is being inserted by a IW plugin. I'll hunt it down
#
[tantek] Bridgy publish urls should only be needed on permalinks that have yet to be POSSEd. If a permalink already has a u-syndication link to a target silo then it should not have a Bridgy Publish link for that silo
#
miklb let me be very clear. I have yet to wrap my head around parsing microformats. So without a validator, I am going on feedback from the community as to what needs to be done. That doesn't seem like an ideal development strategy, but here I am
#
miklb it is a stray insert I believe [tantek] not intended to be there
#
[pfefferle] no offense, I only want to understand what you are exactly doing, to tell you if this causes issues for, for example the semantic-linkbacks plugin…
#
miklb I thought I spent all day yesterday explaining what I was trying to do but I now feel like I've failed at explaining. But since no two people seem to agree what "broken" means this is becoming a more interesting journey
#
miklb it is simple. I want to know what the true limitations in the summer of 2018 of using a default WP theme and the suite of IW plugins, including the mf2 plugin.
#
[pfefferle] can you share your code? perhaps as a github repo? so we have the issues to discuss potential problems…
#
miklb what in that combination "breaks" any other part of the IndieWeb experience.
#
[pfefferle] or you use a branch of the wp-uf2 plugin
#
[pfefferle] broken mf1 implementations
#
miklb I made 1 minor change to the uf2 plugin at this point. I inserted the data u-url
#
miklb that's it. I used the IW plugin h-card widget in the sidebar.
#
miklb but I am working in a branch of that plugin and will send a PR when I'm done.
#
miklb I'm not trying to find fault anywhere, I'm trying to find solutions.
#
miklb or find what absolutely can't be worked around without a change in core. Then tackle that.
#
[pfefferle] should we file the known issues?
#
[pfefferle] to find solutions?
#
miklb sure. some are going to be related to the other IW plugins like the bridgy url one, and some would be the uf2 plugin
[kevinmarks] joined the channel
#
sknebel another thing: the adress part of the h-card is a child, not a property
#
miklb can you be more specific?
#
sknebel in the h-card is a "h-adr" with city and state
#
miklb looks like maybe you opened an issue on that already? https://github.com/indieweb/wordpress-indieweb/issues
#
sknebel that should be "p-adr h-adr"
#
miklb that is coming from one of our plugins.
#
sknebel right, the h-card widget is clear where it is from
#
miklb this is a really good, and I'm going to fix these issues.
#
sknebel miklb++
#
Loqi miklb has 20 karma in this channel over the last year (36 in all channels)
[jgmac1106] joined the channel
#
[jgmac1106] [miklb] [pfefferle] is there a repo somewhere where we can put in the results from each parser....feel like somebody does this every few months, and we never capture it...what would be the best place
#
[jgmac1106] Is this the correct one? indieweb/wordpress-indieweb
#
[jgmac1106] I am happy to repeat your efforts [miklb] now that I can breath again
#
[pfefferle] [jgmac1106] I think there shouldn‘t be parser issues or differences, if so it is a parser bug
#
[pfefferle] So I am not sure we should check this in this context
#
[pfefferle] Miklb why in the indieweb plugin? Shouldn‘t we use the uf2?
jackjamieson joined the channel
#
[pfefferle] miklb, I would not recommend to use the h-card widget, because this is again optional and must be communicated and so a potential error source (can you say that in English?) I would prefer a direct integration into the h-entry context
[manton] joined the channel
#
miklb I'm open to suggestions [pfefferle] I added the widget because it was there and was a stop gap solution.
#
[jgmac1106] [pfefferle] I meant how do we capture the conversation happening here, a native hcard would be nice....but I fear with out plugin you wil bump into jetpack and gravatar messing it up, you need to see what happens then
#
miklb I am not suggesting we are going to provide the most optimal solution with the uf2 plugin. What I am suggesting is with proper documentation and all of the parsing issues on our plugins resolved, we could provide a minimum viable experience with most/any themes
#
miklb there is still going to be room for mf2 themes and educating theme developers about mf2 vs mf1. But if we get all of our code fixed and agreed on what is "good enough", then we can work on that step next.
#
miklb and I'm totally cool doing this live on my site to figure out the state of things.
#
[jgmac1106] awesome
#
miklb but if we are providing a widget that you think we shouldn't be suggesting, that's a problem for us to resolve before going to the theme devs.
#
miklb as to documenting, if we can specify the source of an issue, we should open an issue in the appropriate repo. General discussion has gone in the wp-indieweb repo in the past.
#
[jgmac1106] I also have each of the IndieWeb themes and the top three most popular free themes installed that I am happy to repeat your test [miklb] and give anyone access to those sites
#
[jgmac1106] okay thx for info on general discussion
#
miklb if it is discussion related to specific mf2 that might go in the uf2 repo
#
[jgmac1106] technically for example Independent Publisher is supposed to have a native h-card but I do now gravatar and jetpack can mix it up.....we won't make it anywhere with core if we say "don't use jetpack" so I would say testing conditions require both
#
[jgmac1106] jetpack and gravatar active
[jgarber] joined the channel
#
miklb I'd like to sort out the issues with just the IW plugins first before I introduce Jetpack.
#
[jgmac1106] okay makes sense...just in reference to "we should have native hcard" rather than widget
#
miklb ah, gotcha. That certainly is a discussion that needs to be resolved
#
miklb I need to read up on what native hcard means vs the widget
miklb joined the channel
#
[pfefferle] miklb I hope it is because of my english skills, but I always think we talk past each other
#
miklb there is definitely an inherit risk in text based communication especially when different native languages come into play. I have always tried to account for that.
#
[pfefferle] miklb I do not thing the widget is wrong or bad, I think it is for a different usecase… We talked a lot about usability and I think a user should not have to do things by hand to have a proper mf2 environment… and a widget does not fit all usecases, because we have multi-user blogs and guest posts, … so we should avoid to use a global user as the entry user…
#
[pfefferle] I am still not 100% sure what you are trying to do and why the mf2-feed is not the better solution, but perhaps it might be that I worked on the wp-uf2 plugin for too long…
#
miklb I'm not saying mf2-feed isn't better. It just seems counterintuitive to what semantic html and microformats is all about.
#
miklb as I understand them. So if I can figure out how to get a decent experience with the mf2 plugin and existing IW plugins, then that's what I'm going to do. Until I can't.
#
miklb but parsing issues from our own plugins will be an issue in mf2 feed I believe. So we need to all be on the same page regardless as to what we as a community offer in the plugins.
#
miklb s/regardless/regarding/
#
[pfefferle] not really, because I do not parse the html, I build the JSON manually
#
miklb then keep doing what you are doing and I'll keep doing what I'm doing. We can agree to disagree. I just find issue with one of the core IW plugin developers saying "don't use that widget that is in the community IndieWeb plugin."
#
miklb that is such a mixed message I don't even know where to start.
#
miklb I feel like "perfect is the enemy of good" is coming into play here
#
[pfefferle] ok, that is a good point! I removed it from the IndieWeb plugin and perhaps we should release a new version without the uf2 plugin
#
miklb removed what from the IW plugin?
#
[pfefferle] and re-add either the mf2-feed or wp-uf2 plugins or both if they really work
#
[pfefferle] to keep confusion low
#
[pfefferle] the wp-uf2 plugin
#
[pfefferle] from the recommended ones
#
miklb well, it takes an approval for the PR so you didn't make that decision in a vacuum
#
[pfefferle] no 🙂
#
[pfefferle] but I added the feed plugin, that is also not working yet, so I will also remove that one and release a new version
#
miklb I see.
#
miklb or not. I feel like that is getting me off task.
#
[pfefferle] no, this is also not yet fully functional
#
miklb I'm going to find the source of this bridgy publish being inserted on the home page h-feed.
#
[pfefferle] and I agree with you, we shouldn’t recommend plugins that are not working
#
[pfefferle] to tell users: use the indieweb plugin, but do not install recommendet plugin XY
#
miklb I would have never suggested adding new recommended plugin without fully testing.
#
[pfefferle] then you should perhaps be more verbose on indieweb plugin merge requests 😉
#
miklb you are right. I have been busy the last few weeks and only gave those GH messages a cursory glance.
#
miklb I guess I would never would have assumed anyone would release a new plugin without fully testing it.
#
miklb I have strongly advocated for more beta testing on all of the IW plugins but I should do more to put docs and methods in place to allow more testers
#
[pfefferle] but if you have issues with one of the plugins that is released since ages and you know it casuses issues, please file an issue or make a merge request the next time
#
miklb what are you talking about?
#
miklb I didn't know anything about what was "broken" in mf2 plugin until I installed it last night. So I will most certainly send PR when I find solutions.
#
[pfefferle] we recommend the wp-uf2 plugin in the current version of the indieweb plugin… but as you said, this causes more issues than it helps… so I would vote to remove it until it works
#
miklb I didn't say that. You have been saying that.
#
[pfefferle] > I just find issue with one of the core IW plugin developers saying “don’t use that widget that is in the community IndieWeb plugin.”
#
miklb I want to know what issues it causes and fix them. Either in the plugin or upstream.
#
[pfefferle] ah, you meant me 😄
#
miklb yes, that sends a mixed message.
#
[pfefferle] it really is a language issue
#
miklb so lets find a solution. I'm not looking to point fingers [pfefferle]
#
[pfefferle] I have to read the more carefully
#
[pfefferle] but I have no problems to deprecate plugins if they do not work…
#
miklb if that widget isn't the best solution today, what is, how do we make it so.
#
[pfefferle] and the uf2 plugin really does not work atm..
#
miklb feel free to remove it from the recommended plugins, and I will continue to work on it. Even if it means maintaining a fork outside of the IW repo if there isn't consensus to use it.
#
miklb but right now, the 2 biggest issues that have been pointed out to me about my h-feed have nothing to do with the mf2 plugin, but other IW plugins.
#
miklb 2 of the 3
#
miklb one I have already fixed in the mf2 plugin, the data element with u-url.
#
miklb I can send that PR right now, or I can wait to see if there are other issues I can address.
#
miklb after I fix the other plugin issues.
#
[pfefferle] I have no problem with the plugin and I would love to see a working copy… after all it was initially my idea 😉
#
[pfefferle] the sooner the better!
#
[pfefferle] I prefer a lot of small merge requests over viewer bigger ones
#
miklb I am going to get to the bottom of it. One way or another.
#
miklb OK
tantek, tantek__ and jackjamieson joined the channel
#
miklb time to earn my keep but I'll keep an eye out on any GH issues and will be back this evening
jackjamieson and [asuh] joined the channel
#
[asuh] For beta testing, I think the created GH issue could have an ordered list of steps taken or sites needed to test with. The more points of testing I see that’s needed, the more likely I will be able to follow those directions.
#
[asuh] For micropub and indieauth, I realize there’s a handful of places on the top of my head that I can test against, such as Quill, Indieauth, etc. But to see a given list or other sites I haven’t thought of, or even more ideas of how to use it that didn’t occur to me, is very helpful
[cleverdevil], [deeden] and [jgmac1106] joined the channel
#
GWG If you can post anything using Micropub then that proves it works
[eddie] and [Khurt] joined the channel
#
[Khurt] Wait? uf2 plugin is not recommended?
#
GWG It is too messy
#
GWG miklb, I do want to put this release of Micropub to bed to return to other projects
#
miklb GWG yes, please. Any issues I open right now are purely for documentation and for my own to-do. Feel free to asign them to me or just ignore.
#
miklb [Khurt]: I'm trying to minimize the mess, so consider it a revived work in progress.
#
miklb I would argue mf2 is messy in general
[manton] joined the channel
#
miklb at least to me it feels like there is no pass/fail, more I'll know it is right when I see it.
[pfefferle] joined the channel
#
[pfefferle] aaronpk can I ask, how you optimize your blog posts for mastodon?
#
Zegnat miklb, that is in part because different consumers might look for different data points. There are no hard requirements. Your microformats “fail” when the application you are trying to feed the data to needs more.
[iambismark] joined the channel
#
sknebel eh, there's clear "errors" like things that should be properties being children, properties not having the values intended
#
[pfefferle] aaronpk aaaah, attachments are the magic!
[Khurt] joined the channel
#
[Khurt] Ok. Then I guess I’ll deactivate that plugin. Let’s see what doesn’t work now. My use of IndieWeb tech has been reduced to Webmentions.
tbbrown and gRegorLove joined the channel
#
miklb Zegnat: that sounds like a moving target to me. Which I understand, but I still feel that leaves it a little messy
#
miklb it just would be nice to have a validator that I could pass my h-feed to vs asking folks to look at it and tell me what might be wrong. That's all I'm saying.
#
miklb maybe the validator is based on common application that consume. ¯\_(ツ)_/¯
#
[eddie] It does seem like there could be some type of a validator that could check and confirm that at a minimum. For example: an author, content or summary, a published date. Some “Best practices”
#
[eddie] It also might help just to have an app that “previews” a feed for you. For example test.ltd/preview?ur=https://domain.tld/hfeed/url that would parse the h-feed and then use the parsed info to display a timeline like page. That would allow you to see how a “social reader” might display the feed
#
aaronpk you can basically use xray for that although it returns json instead of displaying the feed
#
aaronpk i keep thinking i should add displaying a preview to monocle since the pieces are almost all there
#
aaronpk btw miklb your posts look better now but are missing the published date now
[grantcodes] joined the channel
#
miklb aaronpk: thanks!
#
miklb a simple validator or preview app would make a good IWC project
[iambismark], globbot, [cleverdevil], [tantek] and [Khurt] joined the channel
#
[Khurt] I saw this on the IndieWeb wiki>
#
[Khurt] > Micro.blog is an indie microblogging service started by Manton Reece, which supports microformats2, Webmention, and Micropub.
#
[Khurt] Why is micro.blog an indie blogging service but Tumblr / WordPress.com is not?
tantek, [eso], tantek__, [jgarber], [shaners], [jgmac1106] and [chrisaldrich] joined the channel
#
[chrisaldrich] [Khurt] In some sense they are, and on some pages they're recommended as such. But they also don't go as far as micro.blog does... each platform/service/CMS/etc. falls on an indie spectrum somewhere
#
[Khurt] [chrisaldrich] ok. But why single out micro.blog in the wiki. It’s just another silos blogging platform (albeit one with IndieWeb tech).
#
[chrisaldrich] there's also a secondary definition/meaning of "Indie" which means not-big, not corporate, etc., so some of it depends on the context/setting.
#
[jgmac1106] [Khurt] if you know of a way to add the plugins to Tumblr/WordPress.com so basic functionality of mf2 parsers and webmentions work we can add it to the wiki
#
[jgmac1106] there are sections on pesos and posse from Tumble and Wordpress.com somewhere I am sure
#
[chrisaldrich] All three are "silos" in some sense in that they are holding your data for you. How easy is it to export your data and move your entire site from one place to another?
#
[Khurt] [chrisaldrich] I’m not sure what corporate means. Micro.blog is a business.
#
aaronpk It's very different from the other models tho
#
[Khurt] [chrisaldrich] how easy is it to export data from WordPress.com?
#
[chrisaldrich] Manton Reece is not at the size or scope of Oath certainly....
#
aaronpk It's not analogous to tumblr or even wordpress.com
#
[jgmac1106] yeah I agree "being the Man" is a wrong filter, and think you can be indie in spirit on Tumblr or WordPress.com
#
[chrisaldrich] They're all apples and oranges on a variety of axes....
#
[Khurt] [aaronpk] how?
#
aaronpk I use micro.blog but none of my content originates there
#
aaronpk It all comes from my site which is my own
#
aaronpk so I use it only as an aggregator and discovery mechanism
#
[Khurt] I use WordPress.com and Tumblr but none of the content originates there, [aaronpk].
#
[chrisaldrich] If you look at what manton is trying to achieve, I think he'd much rather you owned your own domain and your data on your own site, even to the exclusion of you not paying for his service.
#
aaronpk This is always a confusion when new people find micro.blog
#
aaronpk I don't pay for micro.blog hosting
#
aaronpk but I still use micro.blog
#
[chrisaldrich] He's offering micro.blog in some sense as a turnkey for those who don't want to be responsible for the maintenance and overhead for their tech.
#
[Khurt] I don’t WordPress.com or Tumblr for hosting.
#
aaronpk Can you show me how to get my posts from my blog to be followable by people on tumblr? Because I would happily hook that up
#
[Khurt] I know that [chrisaldrich] syndicates his content, in it’s entirety, to WordPress.
#
[chrisaldrich] Like micro.blog, Tumblr and WordPress.com allow you to map your own domain onto their services so that you can take your domain and data with you if you choose to move somewhere else. So in this sense, they're all IndieWeb friendly.
#
[Khurt] https://khurt.tumblr.com
#
aaronpk And by followable I mean not just having me post a copy to tumblr
#
[Khurt] I have followers on Tumblr and WordPress.com
#
aaronpk Manton must have an FAQ that explains this somewhere
#
aaronpk This literally always comes up
#
[chrisaldrich] aaronpk, I think he's admitted in the past that he doesn't have what he considers a good one (yet)
#
aaronpk We should write one on the wiki then :-)
#
[chrisaldrich] I've got a few sketches and some work here on my desk that does a little bit of it already... just haven't circled back yet.
#
[Khurt] I’m just reacting to adding Indie to the micro.blog entry in the Wiki.
#
[Khurt] Seems arbitrary.
#
[chrisaldrich] in some sense, micro.blog can change drastically depending on what tech you're bringing to the table. It's one of the very few services that does it well, but it also means that different people will define it differently
#
[Khurt] micro.blog is sharecropping just like Tumblr and WordPress. Manton controls all of it.
#
aaronpk It's not sharecropping and also we don't use that word anymore
#
[Khurt] Ok. Then let’s call it, “not 100% under my control”.
#
[Khurt] And I don’t understand what’s wrong with sharecropping?
#
[chrisaldrich] [Khurt] Why isn't all of your interaction with micro.blog under your own control? How are _you_ using it?
#
[Khurt] I understand IndieWeb to mean “I make the rules”. Micro.blog has community guidelines. Just like Tumblr, WordPress, etc.
#
[chrisaldrich] All of my content that appears on micro.blog appears on my site first and all the interactions come back to it via webmention.
#
[Khurt] Yes, I do the same, [chrisaldrich].
#
[Khurt] But in the wiki, the word indieweb appears in the context of micro.blog but not Tumblr or WordPress. Why?
#
[chrisaldrich] Perhaps simply that no one has added it directly.
#
[eddie] Because micro.blog is built to intentionally interoperate with IndieWeb sites without any modification
#
[Khurt] here’s the line : > Micro.blog is an indie microblogging service started by Manton Reece, which supports microformats2, Webmention, and Micropub.
#
[Khurt] https://indieweb.org/Micro.blog
#
[eddie] You have to customize, tweak and otherwise coerce Tumblr and Wordpress to interoperate with our websites
#
[Khurt] Would it be ok to add the same word, “indie”, to the wiki entry for tumblr.
#
[chrisaldrich] At one time, I experimented with using Tumblr as my online identity. They allowed easy import/export of data and I could map my domain onto them. Because I can control the theme I could also add microfomats too.
#
[chrisaldrich] It's even possible to hook Brid.gy up to tumblr... so in all those senses it's very IndieWeb.
#
[chrisaldrich] Are many people using it that way? Probably not.
#
[Khurt] [chrisaldrich] that is what I am getting at.
#
[chrisaldrich] I suspect that you wouldn't see any torches or pitchforks for doing so.
#
[Khurt] here’s what the wiki entry for tumblr: “Tumblr is a popular content hosting silo, ”
#
[chrisaldrich] I think that's a result of the fact that that's the way 98% of people have traditionally used Tumblr.
#
[Khurt] So how is Tumblr a silo but micro.blog is “indie”?
#
[Khurt] Politicis.
#
[chrisaldrich] Most of the people and companies that were using Tumblr as a personal or corporate blogging platform jumped ship about 2 years ago and moved to Medium.com when that was in vogue.
#
aaronpk You do know this is a wiki that anyone can edit right? You're welcome to change it as you see fit
#
[Khurt] I think the wiki entries are using words to give the impression that Tumblr is a silo but micro.blog is not.
#
[Khurt] [aaronpk] will do so.
#
[chrisaldrich] Some of those politics are also tied up in what the owning corporation is actively doing with their platform.
#
aaronpk Chances are that wiki page was created a long time ago, before micro.blog was a thing
#
[chrisaldrich] Oath/Yahoo! hasn't been showing much (any?) love to Tumblr for quite some time.
#
aaronpk Nobody looks at all the pages on the wiki and says "what are the most effective words to use to push our agenda between these four companies"
#
[Khurt] to be fair, I think Tumblr and micro.blog are both silos that can be used to build indie web site. Although I like to think of IndieWeb as 100% independent (aka. under my control from server stack to web stack).
#
aaronpk Turns out that not everyone has that same definition tho
#
[Khurt] [aaronpk]yeah unfortunately.
#
[chrisaldrich] [Khurt], yes there's that, but we also can't and don't expect "everyone" to make that their go to definition. To reach the so-called 3rd and 4th gen users, there are going to have to be companies like micro.blog, Tumblr, and WordPress.com do some of the (heavy) lifting for them.
#
miklb saw all the mention of Tumblr and and micro.blog and thought I was in a different channel
#
[eddie] Lol yeah we probably should have shifted
#
[chrisaldrich] Ideally, with more openness and support of standards, then pure competition takes over and these companies will be offing more and hopefully cheaper services and solutions
#
[Khurt] [chrisaldrich] yep. [miklb] it’s all related.
#
[Khurt] [chrisaldrich] FYI: I supported the micro.blog kickstarter. Was hoping I’d get to run my own server.
#
[chrisaldrich] then shall we move to --> #indieweb channel?
#
miklb I was just joking around. I haven't figured out how to have more than one channel in irssi so the more the merrier :-)
#
[chrisaldrich] I'm happy to stay here, particularly if it means we lose participants....
#
tantek__ whoa non-wordpress scrollback!
#
miklb what is a representative h-card
#
Loqi The representative h-card for a page is an h-card on that page that represents that page, if any, as not all pages are about a person or organization, a page might not have a representative h-card https://indieweb.org/representative_h-card
#
[eddie] Yeah we needed chrisaldrich to poke us into the proper channel
#
[eddie] oops miklb
#
miklb an h-card on that page that represents that page ???
#
aaronpk That's not very well worded but it is accurate if you stare at it long enough
#
tantek__ feel free to suggest wording improvements - I think it's a particularly hard thing to express
#
[Khurt] It works?
#
[Khurt] That should be “It works!!!”
#
miklb I'm just looking at https://github.com/indieweb/wordpress-indieweb/issues/113 and wasn't familiar with the term. Not sure I could suggest better phrasing at this point, but it is awkward.
#
Loqi [sknebel] #113 provide representative h-card
#
GWG Didn't I fix that?
#
miklb fix what?
#
GWG The missing u-url
#
miklb in the uf2 plugin?
#
miklb needs more information
#
miklb ACTION 
#
GWG In the indieweb plugin
#
miklb dunno. Issue is still open so I'm just reading through all issues related to h-card
#
GWG I will have to look at the fixes from two weeks ago over breakfast
#
miklb there was a parsing issue in the addr mentioned this morning. Going to tackle the h-card tonight.
#
miklb if you have code you haven't pushed or issues that didn't get closed let me know if you get a sec
#
GWG Commented
#
miklb I honestly think if we iron out a few confusing pieces and tweak a few things between the existing plugins, the mf2 plugin experience with the other IW plugins isn't bad
#
miklb I am not saying its a drop in replacement yet
#
miklb somewhere I recall a conversation about h-entry on the body class a post permalink, and if that was still the right way to go. Anyone remember it?
#
miklb that might be tantek__ realm of conversation https://github.com/indieweb/wordpress-uf2/issues/14
#
Loqi [dshanske] #14 hentry
[pfefferle] joined the channel
#
[pfefferle] Yes
#
[pfefferle] I added the h-entry to the body, to support comments
#
[pfefferle] So we had to move also the hentry, because parsers had problems with hentry on a different HTML element than h-entry if backwards compatibility was enabled
#
[pfefferle] Might have changed now
#
[pfefferle] I can say that it is the right way to add the h-entry to the body, does anyone know if there are still problems when hentry is on a different tag than h-entry? /cc Zegnat [gregorlove]
#
sknebel yes, nothing in microformats has changed in that regard
#
tantek [pfefferle]: yes for new pages / sites, primary microformat on the body makes sense. However for such a complex setup / history / context as WordPress, not sure if that suggestion helps at all
#
[pfefferle] it will also work, if we do not have to put the microformats2 classes on the same element as the microformats1 classes, do you know if this is currently possible with php-mf2 [tantek]?
#
tantek it's possible yes, however you will get nesting
#
[pfefferle] nesting?
#
[pfefferle] between both versions?
#
tantek yes due to backcompat handling
#
sknebel the parsers parse both mf1 and mf2
#
sknebel so you can put mf1 objects in mf2 and vice versa
#
tantek right, so you can add just one mf2 object to an existing structure and everything else keeps working
#
sknebel but if you want to mark up one object with both, you need to put the classes on the same html elements
#
tantek correct
#
[pfefferle] miklb that makes things even worse, because that means that we cannot add a date or a title
#
sknebel <div class="h-entry hentry"> is parsed as an mf2 h-entry, since the h-entry takes priority
#
[pfefferle] at least not for themes that are already supporting microformats1
#
[pfefferle] thanks tantek and sknebel
#
sknebel <div class="h-entry"><div class="hentry"> parses as two nested h-entries
#
sknebel [pfefferle]: that's only for h-* classes!
#
[pfefferle] ah, not for properties?
#
sknebel if inside a h-* you have a class name that's a mf1 property name, that gets ignored
#
[pfefferle] ah, that’s nice
#
tantek yes that's deliberate because it simplifies the authoring model to something more predictable / maintainable per object
#
[pfefferle] sure
#
[pfefferle] ok, but then we still have the h-feed issue
#
[pfefferle] because hfeed is hardcoded in the template and we add h-feed to the body
#
[pfefferle] so we can’t add h-feed with the mf2 plugin
#
[pfefferle] and we have to set the h-card to the post container, that sadly means no comments support
#
sknebel hfeed on the feed pages only?
#
[pfefferle] no, on every page
#
[pfefferle] but right now, we also add h-feed and hfeed to the body
#
[pfefferle] so we have 3 h(-)feeds in the site
#
[pfefferle] and that causes http://pin13.net/mf2/?url=https%3A%2F%2Fichmagdiesen.link%2F
#
[pfefferle] the nested problem
#
[pfefferle] sknebel++
#
Loqi sknebel has 5 karma in this channel over the last year (80 in all channels)
#
sknebel yeah, that's a problem then with themes that do that then :/
#
[pfefferle] tantek++
#
Loqi tantek has 3 karma in this channel over the last year (93 in all channels)
#
tantek is there really no way to have a plug-in do one last pass on the output coming from the theme before it is returned in the HTTP request?
#
[pfefferle] I would vote to skip the uf2 plugin on themes that support mf1…
#
tantek [pfefferle]: that seems like a reasonable approach to try
#
[pfefferle] backwards compatibility parsing seems better than the mixture of both
#
[pfefferle] but it is not that easy to check if a theme supports mf1
#
[pfefferle] 😞
#
tantek that does seem to be what is happening in this situation, plus we can improve backcompat more!
#
tantek if we do that, then I need your (and others) help with suggestions here: https://github.com/microformats/h-entry/issues/7
#
Loqi [tantek] #7 Can we improve h-entry backcompat to handle common WordPress hentry streams?
#
sknebel theme with random hfeed on every page will parse oddly with backcompat too though?
#
[pfefferle] yes, but I handle this on the parser level
#
sknebel so it works with other wordpress sites
#
[pfefferle] yes
#
[pfefferle] if the rest of the mf1 is properly used
#
sknebel you also loose all the properties only in mf2: replies, likes, bookmarks, ...
#
[pfefferle] sure, that’s the risk
#
[pfefferle] and we are again at a level, where I vote for the mf2-feed idea 😉
#
sknebel I totally get why you want to do that
#
tantek__ [pfefferle] would really appreciate your opinion at least on the current proposals in that issue 7
#
tantek__ if we can improve backcompat parsing for a massive % of WordPress sites by default with just a few simple class/rel rules, that would be a very big win
#
[pfefferle] Sure, will try to match this with our theme problem
#
[pfefferle] A big win would be if we can combine mf1 containers with mf2 properties
#
tantek interesting
#
[pfefferle] In that case we have to only add mf2 properties and it would work with the old mf1 and the new mf2 themes without changing „core“ semantics
#
tantek pfefferle - what about the things that are already marked up but with other (non-mf1) class names?
[eddie] joined the channel
#
[pfefferle] Do you have an example?
#
tantek yes the two in issue 7 :)
#
[pfefferle] If we can add a p-author to an hentry this would be no problem
#
[pfefferle] I think this should be possible through a plugin
#
[pfefferle] If something isn‘t marked up, it is not that much of a problem
#
[pfefferle] The bigger problem is for example the entry-title... it is marked up as mf1 but has no hooks to add mf2 properly
#
[pfefferle] To be more specific, the nicest way would be to have mf1 containers that are supporting mf1 and mf2 properties
#
[pfefferle] So that we can re-use hentry, entry-title, entry-content and published/updated and add p-author h-card and the other metas through the plugins
#
tantek alternatively we explicitly add a small # of mf2 classes as part of the backcompat handling for h-entry
#
tantek instead of making all mf1 classes handle all mf2 classes which is a pretty big change
#
[pfefferle] Hmm, but GWG uses a wide range of formats with post-kinds, I think...
#
GWG As many as people ask for.
#
[pfefferle] This is a deadlock 😝
#
tantek pfefferle, all those properties are in h-entry
[kevinmarks] joined the channel
#
[pfefferle] We could add an h-entry, and an e-content, but it is hard (or impossible) to add a p-name
#
[pfefferle] All other properties like date, permalinks and authors has to be added as hidden data
#
[pfefferle] If, we could add p-name as hidden data
#
[pfefferle] But that would be against the microformats principle...
#
tantek my goal for issue 7 was to at least pick-up more date properties automatically as well as author
#
tantek and any others that are already consistently marked up across themes
#
[pfefferle] I will have a look at some themes, to give you some feedback tomorrow
#
[pfefferle] Too late today
jackjamieson and davidmead joined the channel
#
[pfefferle] [tantek] I don‘t think we have to do so many backwards compatibility things... the mf1 stuff seems to work pretty well http://pin13.net/mf2/?url=https%3A%2F%2F2017.wordpress.net%2F2016%2F10%2F17%2Fnow-offering-square-one-coffee%2F
#
Loqi [Needle & Hook] Now offering Square One coffee
#
[pfefferle] This is the classic 2017 theme from Wordpress
#
tantek agreed, and hence why I opened issue 7 - to focus on the few remaining things to see if we can get those last few properties :)