#dev 2018-05-04

2018-05-04 UTC
[miklb], [chrisaldrich], tantek, eli_oat, steph1, renem, leg, [kevinmarks] and gRegorLove joined the channel
#
GWG
aaronpk: Did as you asked. Waiting for review
#
Loqi
gwg has 30 karma in this channel (339 overall)
KartikPrabhu, cweiske, [unoabraham], [jeremycherfas], loicm, [chrisaldrich], iasai_, AngeloGladding, swentel, jeremycherfas and [pfefferle] joined the channel
#
klezlab.it
edited /IRC_People (+78) "/* Nicknames */ added myself"
(view diff)
[colinwalker] joined the channel
#
Zegnat
is struggling with p3k-http and should just consider not using aaronpk’s code for *everything*
[jgmac1106] joined the channel
#
Zegnat
Happy to say that my code for having PHP keep track of the redirects curl follows seems to be working though :)
#
martymcgui.re
edited /Morris (+1683) "add details, user list, features, source code link"
(view diff)
#
skippy
ancarda: fail2ban can be configured to read from pretty much any file. different jails can have different log sources. My ssh jail reads from syslog; while my WP jail reads from a file in my WP directory space.
[sebsel] and barpthewire joined the channel
#
aaronpk
Oh does it know the difference between 302 and 301? That was a trick
[jgmac1106] joined the channel
#
Zegnat
The redirects array will contain tuples of http code and origin url
#
Zegnat
And no need to count or loop redirects now, as it'll just use curls max redirects.
leg joined the channel
#
Zegnat
Hmm, maybe I should do a dictionary with code/from/to instead, like you do there!
#
Zegnat
aaronpk, interested in this as a PR later?
#
aaronpk
Yea that'd be awesome! I couldn't figure out how to do that with the curl functions quickly so I did it in guzzle instead
#
aaronpk
I don't like how guzzle throws exceptions when it encounters errors tho, especially since the list of exceptions doesn't seem to be documented anywhere
#
Zegnat
Exceptions for HTTP errors or transport errors? I wouldn't expect them for http error codes, as those are part of a valid response.
#
Zegnat
Either way, I'll see if I can update the redirects array to match your guzzle one, and then send a PR your way!
snarfed and [pfefferle] joined the channel
#
skippy
https://indieweb.org/reply-context has the quoted text in a <p>. why do that, instead of a <blockquote> ?
#
skippy
also, what's the difference between "class=p-content" and "class=e-content" ?
#
Zegnat
1: no reason, just how the example ended up. Blockquote would have been nicer.
#
Zegnat
2: p-* results in plain text only, while e-* instructs the mf2 parser to also return the exact HTML
#
skippy
thanks Zegnat
#
Zegnat
So e-* often makes sense for content. You’d want to give readers the full markup, not just plain text.
#
aaronpk
i can't find a link for that question but it comes up all the time
#
Zegnat
Maybe the next wiki page I create is just a non-parser-implementer guide to mf2 prefixes, haha
#
skippy
http://microformats.org/wiki/h-entry describes e-content, but not in the way that Zegnat just did.
#
Loqi
[Tantek Çelik] h-entry is a simple, open format for episodic or datestamped content on the web. h-entry is often used with content intended to be syndicated, e.g. blog posts. h-entry is one of several open microformat standards suitable for embedding data in HTML. ...
#
Zegnat
I don’t see a description there? I only see the general function of the content property described there (it being “the full content of the entry”)
#
skippy
that's more description than I found anywhere else.
#
skippy
so maybe s/describes/mentions/ is better in my previous comment.
#
Loqi
microformats 2 prefix conventions
#
aaronpk
Tho that could be improved a bit, it's pretty jargon heavy
#
skippy
thanks for the link and explanations!
tantek joined the channel
#
Zegnat
aaronpk, after some tweaking, the redirect list straight from p3k-http: https://gist.github.com/Zegnat/2659a7a6fb40398aef03e8c64035e9e0
#
Zegnat
I forgot what domain we said had a 302..?
#
Zegnat
p83.nl :) And it shows a 302
maingo joined the channel
#
Zegnat
aaronpk, if you have any proper redirect testing for indielogin.com, feel free to grab this and give it a go: https://github.com/aaronpk/p3k-http/compare/master...Zegnat:curl-redirects
#
Zegnat
I have to run for dinner, will write a PR tonight
#
aaronpk
Thanks! I have a bunch of test identities that try out all the redirect possibilities but haven't published them yet
#
aaronpk
I'm hoping to eventually put them at IndieAuth.rocks
tantek and [eddie] joined the channel
#
[eddie]
aaronpk: Have you tried logging into the micro.blog macOS app? I’m having issues, I think it’s related to the expecting me returned issue that Sunlit had before. But not 100% sure so I was curious if it was working for you
#
aaronpk
I haven't tried yet
#
[eddie]
Gotcha :thumbsup:
#
[eddie]
Ohh you almost might run into the issue that if it detects an xml-rpc header it tries to log in that way
AngeloGladding joined the channel
#
[eddie]
I sent both issues to manton, so I’ll let you know the verdict when I hear back
#
GWG
[eddie]: Got a moment?
#
tantek
Zegnat, and anyone else asking about my single_emoji heuristic, I've added it to the CASSIS projecti in a new file for experimental functions: https://github.com/tantek/cassis/blob/master/cassis-lab.php (though experimental, both functions are live deployed on my site as part of /Falcon )
#
[eddie]
hey GWG, yeah I do
#
GWG
[eddie]: I saw you mention somewhere you were using the micro.blog API. What for?
#
[eddie]
Yeah, I’ve been using the micro.blog api to fetch mentions and push them into my Aperture server
#
[eddie]
for notifications
#
GWG
Interesting.
#
tantek
spoiler: detecting a single emoji in only ~25 lines of PHP
#
GWG
I seem to be seeing a lot of interest in the WordPress Indieweb ecosystem from micro.blog users and trying to think how I might entice them
#
GWG
Gateway drug type of idea
#
tantek
GWG, anything you can do to remove struggle, remove pain. make it as painless as possible. keep on simplifying UX, reducing total # of plugins.
#
[eddie]
Ohhh interesting.
#
[eddie]
Yeah I agree with tantek. One big thing with people on micro.blog is the easy of entry. The biggest barrier from micro.blog people joining the greater indieweb is the increased complexity from the conversations I’ve seen
#
tantek
GWG, I'll give more feedback in #indieweb-wordpress
#
eddiehinkle.com
edited /Webmention-faq (+779) "Added questions from a blog post"
(view diff)
#
tantek
also a cassis.js update with a couple more URL parsing functions (for path segments and fragments)
#
GWG
aaronpk, by the way, pfefferle pushed those Indieauth changes, so tell me if OYS or such are still getting issues filed
#
aaronpk
oh great! I won't know until people start updating their websites with it
#
aaronpk
is it on wordpress.org now
#
GWG
aaronpk, that's where he pushed them
#
GWG
He mentioned that the Auto deployment script is working, so I may do the next one
#
GWG
I can go back to my other project till another bug is filed
#
eddiehinkle.com
edited /Webmention-faq (+364) "/* Do webmentions work only when both sites use webmentions? */ initial pass at answering this question"
(view diff)
kaushalmodi joined the channel
#
kaushalmodi
eddiehinkle: I understand the part where the sending website need not have mf2 markup.
#
kaushalmodi
I am might be wrong, but shouldn't the receiving website *have to* support webmention, but specifying the endpoint in the rel=webmention?
#
[eddie]
Yeah, that’s what I said, isn’t it? for the receiving website to have a webmention endpoint
#
kaushalmodi
If the receiving website doesn't have the endpoint, where will that webmention be logged? From what I understand, it has to be captured by some endpoint, right?
#
kaushalmodi
wait, let me re-read
#
kaushalmodi
Ah OK, this statement confused me: "both sites don't have to support sending and receiving webmentions"
#
GWG
Minimum viable is a webmention form anyone can use
#
[eddie]
That could be tweaked and/or removed
#
kaushalmodi
I think you mean: "both sites don't have to support *both* sending and receiving webmentions" ?
#
kaushalmodi
.. that another 'both'
#
[eddie]
Yeah, that’s what it is
#
kaushalmodi
Also the sending website has no requirements
#
kaushalmodi
So this part is not necessary: "The minimum requirement for a webmention exchange to work is for sending website to support microformats"
#
kaushalmodi
If sending website has mf2, the receiving website can better understand the webmention..
#
kaushalmodi
like is this a reply, or a like, or a report, or a ..
#
[eddie]
Hmmm, yeah I guess a webmention with mf2 is just a “mention”?
#
kaushalmodi
But if the sending website has no mf2, it appears just as a link
#
[eddie]
gotcha.
#
kaushalmodi
the mf2 in sending website is to help the receiving website better format that wm
#
tantek
better present
#
eddiehinkle.com
edited /Webmention-faq (+156) "/* Do webmentions work only when both sites use webmentions? */ some tweaks based on irc discussions"
(view diff)
#
[eddie]
How about that?
#
[eddie]
Feel free to tweak if you all see more areas of improvement 🙂
#
kaushalmodi
Looks good
#
kaushalmodi
Just a question, as this is out of my webmention-fu, does the sending website even need the link of the receiving website?
#
kaushalmodi
I mean, what happens if someone uses the "Send Webmention" form on the receiving website and enters a link that does not contain a reference to that receiving website?
#
[eddie]
From what I understand, most webmention implementations should reject the webmention
#
kaushalmodi
That makes sense if that's the case; just that that question never came to me till now :)
#
[eddie]
Yeah, I think it is considered mandatory because the spec says “MUST perform a HTTP GET request on source, and follow any HTTP redirects (up to a self-imposed limit such as 20) to confirm that it actually links to target.”
#
eddiehinkle.com
edited /Webmention-faq (+381) "/* How do you know if a website uses webmentions? */ initial pass at answering question"
(view diff)
#
[eddie]
Feel free to critique that one as well. I have to run and give my cat her meds. be back
#
[eddie]
(back)
#
scripter.co
edited /Webmention-faq (+39) "Minor reformatting of "Do webmentions work only when both sites use webmentions?""
(view diff)
#
kaushalmodi
Hope that edit's alright, and correct,
#
[eddie]
perfect!
#
tantek
[eddie]++
#
Loqi
eddie has 31 karma in this channel (56 overall)
#
tantek
kaushalmodi++
#
Loqi
kaushalmodi has 3 karma in this channel (4 overall)
#
eddiehinkle.com
edited /Webmention-faq (+181) "/* If I write an article that links to another article and both websites support webmentions, does it notify the other website automatically? */ Initial pass at answering"
(view diff)
#
eddiehinkle.com
edited /Webmention-faq (+310) "/* How do webmentions work with social media sites like Twitter? */ initial pass at answering"
(view diff)
#
[eddie]
Alright, that’s all the answers I’ve got in me 🙂 Anyone else, feel free to tweak and/or answer the other questions
KartikPrabhu, snarfed, renem and [pfefferle] joined the channel
#
Zegnat
Oooh, FAQ work, interesting!
tantek and KartikPrabhu joined the channel
#
martymcgui.re
edited /Netflix (+357) "/* How to export */ add my example(s) and write-up blog post"
(view diff)
#
martymcgui.re
edited /Netflix (+257) "/* Ratings */ more scraping"
(view diff)
#
aaronpk
Zegnat: re p3k-http, is your change backwards compatible, and also is it at least not going to make the curl vs file_get_contents transports out of sync?
#
skippy
another question about reply contexts. Is it valid to have soemthing like <div h-entry><div e-content><div u-in-reply-to h-cite><blockquote p-content>quoted content</blockquote></div>my original remarks</div> ?
#
skippy
or will parses not like the quoted bits inside the e-content ?
#
Zegnat
Aah. Hmm. I guess the other transports (e.g. stream) won’t have the redirects property added to their output.
#
Zegnat
Other than that, it shouldn’t have changed anything. Other than the debug property being even more complete now (containing all headers from all redirect steps)
#
aaronpk
skippy: you'd want your own text in the e-content, and the reply context outside of the e-content
#
Zegnat
The idea was that it really only adds the redirects property, aaronpk. I am not sure what to put there for other Transport mechanisms? Would you expect `null` from Stream?
#
aaronpk
Zegnat: yeah i guess if it's just adding a property you can leave it out of the others
#
Zegnat
All it does. It collects headers differently too. But it then turns those into a string again for the overlaying Http class to reparse into arrays again. Just to stay backwards compatible ;)
#
Zegnat
(Part of the “wrangling” I mentioned this morning.)
snarfed, tantek, KartikPrabhu, renem and [chrisaldrich] joined the channel
#
aaronparecki.com
edited /GDPR (+162) "move tweet from /email"
(view diff)
#
aaronparecki.com
edited /email (-162) "move tweet to /GDPR"
(view diff)
[kevinmarks] and gRegorLove joined the channel
#
www.boffosocko.com
edited /Webmention-faq (+490) "/* How do you know if a website uses webmentions? */ practical advice as well as mention of buttons"
(view diff)
#
www.boffosocko.com
edited /Webmention-faq (+659) "moved non-dev FAQ up the page; notes on common sites which will automatically send webmentions"
(view diff)
#
www.boffosocko.com
edited /Webmention-faq (+1027) "stab at an answer for setting up webmentions for Gen2+"
(view diff)
iasai joined the channel
#
jmac
I'm not clear on how one is meant to render a (manually submitted) webmention whose source document is Just Some Webpage that's devoid of MF2
#
aaronpk
I show just the link when that happens, with a placeholder avatar
#
jmac
Do you just print something like "Here is a mention: [title of page]" with no attempt to tease out content?
swentel joined the channel
#
aaronpk
I could probably improve mine a bit since right now it's just like my normal comments but missing all the content
#
GWG
How do you pixelate an image? Is there a library?
chrisaldrich joined the channel
#
swentel
oh stackoverflow has answers probably :)
#
Loqi
swentel: tantek left you a message 2 days ago: any better now? https://twitter.com/swentel/status/984772721337929729 :)
#
tantek
oh boy
#
swentel
yeah, it's better :)
#
tantek
sgreger has demonstrated that 3x3 works well. I wonder if 4x4 is still doable
#
swentel
I wonder, is he using webmention.io ? because it's annoying if you need to start pulling those images to your own server as well to pixelate them ..
#
tantek.com
edited /Webmention-faq (+754) "/* FAQ */ update webmentions work only when both sites a bit with more details, explicitly why microformats2"
(view diff)
#
sknebel
swentel: No, he modified his WordPress
#
tantek
what is pixelate
#
Loqi
It looks like we don't have a page for "pixelate" yet. Would you like to create it? (Or just say "pixelate is ____", a sentence describing the term)
#
sknebel
In the announcement post for that feature, he even might have code, not sure
#
tantek
worth creating a stub for
#
swentel
thinking to add anonimyze options too in the drupal plugin
#
swentel
but probably rather use some sort of anonymous image
#
swentel
instead of pulling them in myself
#
swentel
anonymize
#
swentel
(or however you write that damn :)
[pfefferle], [chrisaldrich], [tantek], [miklb], tantek, [unoabraham] and snarfed joined the channel
#
jmac
Followup from my griping yesterday about tumblr, and how it makes manual webmentions tricky because of its broken use of internal pseudo-redirects to external sites:
#
jmac
How would we feel about a Webmention implementation marking a source document that has the target URL as verified, even if that URL is URI-escaped (a.k.a. percent-encoded)?
#
jmac
snarfed suggested to me that my implementation just have a tumblr-specific exception that looks for a "z=" query-string parameter and uses that. But if the webmention parser can generally allow for URI-escaped version of the target URL, then that'll also work without having to resort to a yucky tumblr exception