#dev 2017-07-24

2017-07-24 UTC
j12t, eli_oat, KartikPrabhu, j12t_, Defenestrate and gRegorLove joined the channel
#
gRegorLove
!tell schmarty Not sure if consuming software gripes about this, but apparently iCalendar standard says you can't use UTC offset in date-time values. Noticed your .ics files have them.
#
Loqi
Ok, I'll tell them that when I see them next
#
gRegorLove
Appears it can be of form 1) date with local time 2) date with UTC time or 3) date with local time and timezone reference
#
gRegorLove
Latter being the fun stuff like "TZID=America/New_York"
#
gRegorLove
I'm going with 2 on this project, heh
cweiske and barpthewire joined the channel
#
loqi.me
created /perma.cc (+57) "prompted by cweiske and dfn added by cweiske"
(view diff)
#
gregorlove.com
edited /perma.cc (+145) "expand dfn, links"
(view diff)
[pfefferle] joined the channel
#
[pfefferle]
good morning
#
Loqi
*yawn*
j12t joined the channel
#
gRegorLove
Does anyone have examples of h-event published with u-uid?
#
gRegorLove
Specifically where it's different than the u-url
j12t joined the channel
#
@DavidDarnes
Webmentions seem pretty cool for article commentary, I wonder if it's possible to integrate this now with a Jekyll… https://twitter.com/i/web/status/889429859142955008
(twitter.com/_/status/889429859142955008)
davidmead and KevinMarks joined the channel
j12t, jonnybarnes, Defenestrate, KevinMarks and KevinMarks_ joined the channel
#
schmarty
!tell gRegorLove thanks! My .ics files are definitely not to spec but Google calendar seemed to grok them so good enough for now. Date with UTC time would be my preference.
#
Loqi
Ok, I'll tell them that when I see them next
#
Loqi
schmarty: gRegorLove left you a message 6 hours, 17 minutes ago: Not sure if consuming software gripes about this, but apparently iCalendar standard says you can't use UTC offset in date-time values. Noticed your .ics files have them.
KevinMarks, j12t, j12t_, KevinMarks_, singpolyma, eli_oat and snarfed joined the channel
#
petermolnar
if this actually works with random aspect ratios, I'll be very excited in a few minutes: http://www.cssscript.com/responsive-justified-photo-grid-layout-pure-css-flexbin-css/
#
Loqi
Responsive Justified Photo Grid Layout With Pure CSS – flexbin.css
#
petermolnar
eh, same problem as before, it crops
#
prtksxna
petermolnar: Are you ok with using js?
#
petermolnar
I know this can be done with js :)
#
petermolnar
I'm stunned that flexbox can't do anything similar
#
petermolnar
I've asked around a few months ago, I though this is some new approach; it's not
#
prtksxna
hasn't explored flexbox/css grid enough to try hacking on this
#
prtksxna
will soon
#
prtksxna
petermolnar: You'd be barely compatible with browsers though :/
#
petermolnar
still doesn't work without any crop
#
prtksxna
petermolnar: Oh... what exactly is your use case?
#
aaronpk
oh yeah you want the flickr style one right/
#
Loqi
aaronpk: tantek left you a message 1 day, 16 hours ago: question about the federation support demonstrated by https://aaronparecki.com/2017/07/19/24/webmention-io - when you're looking at that post yourself when signed-into your site, do you have/see "Reply" or "Like" buttons yourself on the comments? And if you reply right there, do they get threaded with the original comment on micro.blog?
#
petermolnar
but it's not something I can't live without
#
aaronpk
yeah that's a trick
#
aaronpk
petermolnar: i think it's possible to do without JS if you know the image dimensions on the server side and do the layout there. is that not an option for you?
#
petermolnar
in that case, it'd hard to keep it fluid unless you pre-organize the whole thing on the server side
#
aaronpk
oh fluid too? yeah no
#
petermolnar
which, right now, is just too much effort to do for me :)
#
petermolnar
yep, that is the BIG trick in it
#
aaronpk
in order to do that layout fluid, you have to recalculate the whole thing each time
#
petermolnar
so I thought that with flexbox and max vw/vh you can do it
#
petermolnar
hm, only one of those, this is a really nice comparison
#
petermolnar
thank for the lin
#
petermolnar
but yeah, so vw / vh as units is how I keep my images fluid and locked aspect ratio right now
#
petermolnar
and it's working nice
#
petermolnar
but flexbox is much more limited than I anticipated
#
petermolnar
I miss flickr from that comparison
#
Loqi
misses flickr from that comparison too
KartikPrabhu and jeremycherfas joined the channel
davidmead joined the channel
#
jeremycherfas
Hello zegnat
#
jeremycherfas
I’ve run into a strange problem with the continuing work on Grav webmentions.
#
jeremycherfas
I updated the plugin to the 1.1 release, and that created a fatal error.
#
jeremycherfas
If I comment out just that line, I no longer get the error.
#
jeremycherfas
With that line, Symfony says “Unexpected characters near "]" at line 40 (near "links: { }]”).” and tracing the various steps in Grav it goes back to that one line.
#
jeremycherfas
But I cannot see anything funny in that line.
KevinMarks_ joined the channel
#
voxpelli
must say that he loves that he built language sniffing into his Micropub endpoint – he just posted a bookmark in Swedish and it automatically understood it was Swedish and marked it up as Swedish – yay
#
Loqi
does a happy dance!
#
voxpelli
hugs Loqi
#
Loqi
hugs voxpelli
dougbeal|mb1 joined the channel
#
Zegnat
jeremycherfas, that error suggests a problem in the YAML file to me
#
jeremycherfas
Webmention.yaml?
j12t joined the channel
#
Zegnat
possibly.
#
Zegnat
Does that have a “links” near line 40?
#
jeremycherfas
No, that contains details for the Vouch settings.
#
jeremycherfas
I could try disabling Vouch
#
jeremycherfas
That makes no difference.
#
jeremycherfas
Let me try creating an entirely new YAML file.
#
jeremycherfas
Nope, same error.
#
jeremycherfas
New, clean YAML is only 32 lines, but get the same Line 40 error.
#
jeremycherfas
That line 40 is in Symfony’s /yaml/inline.php file and is the line that produces the error message.
#
jeremycherfas
It might be something to do with the datafile that the plugin creates.
#
jeremycherfas
The plugin is calling Yaml::parse on the content of the datafile
#
jeremycherfas
Yup. That was it. I deleted the old datafiles and the error is gone.
#
Zegnat
Yeah, had to be YAML :/
#
Zegnat
Wonder why though, because I believe it uses the same YAML lib to write the file in the first place
snarfed joined the channel
#
jeremycherfas
Possibly because the files were left over from a previous install. I got in a terrible mess yesterday, so deleted the Dev branch and re-installed the plugin, now working on a new dev branch and the first thing that happened was that error. :(
j12t joined the channel
#
Zegnat
Hmm, odd.
#
jeremycherfas
Probably an error I introduced.
#
Zegnat
It is really hard to mess up saving and reading the YAML file though, jeremycherfas :P
#
jeremycherfas
Not when you open it to read and maybe press a key by accident.
#
Zegnat
Ah, well, yes, that could be it
#
Zegnat
These days I am a big fan of opening such files on the command line where I am sure not to accidentally save them
#
jeremycherfas
That makes sense.
#
jeremycherfas
Voxpelli Is it possible to use your webmention app from a local install?
#
voxpelli
jeremycherfas: I run it all locally in dev, with my testpinger and everything – just check the sample env file and then run the command to set up the local Postgres database
#
voxpelli
You of course need to install Postgres then as well :) But that should be easy, depending on your OS
#
jeremycherfas
Voxpelli: I might try and do it by exposing my local dev site with ngrok.
#
voxpelli
That works as well
tantek joined the channel
#
jeremycherfas
Simpler, for me, to begin with.
#
voxpelli
jeremycherfas: if you want Postgres on macOS: https://postgresapp.com/
#
voxpelli
Should work well, I do mine via Homebree though
#
voxpelli
s/Homebree/Homebrew/
#
jeremycherfas
I generally use Homebrew too.
snarfed, j12t_ and gRegorLove joined the channel
#
gRegorLove
I've got an MVP PHP script to make .ics from h-event working. Will open source it shortly once I clean it up.
#
Loqi
gRegorLove: schmarty left you a message 5 hours, 4 minutes ago: thanks! My .ics files are definitely not to spec but Google calendar seemed to grok them so good enough for now. Date with UTC time would be my preference.
#
tantek
gRegorLove: by "working" I seriously hope you looked at how H2VX generates ics and copied the weird details in the ICS output from it
#
tantek
otherwise it's going to be many months of "fix the ICS to work in all the calendar consumers one at a time" struggles
#
gRegorLove
Are the weird details documented or just by reading the xsl?
#
gRegorLove
I followed the RFC and the validator on icalendar.org, as well as comparing to h2vx output and schmarty's
#
tantek
I think the weird details may be documented in ancient X2V commit logs
#
tantek
yes the RFC + validator are very far from sufficient
#
tantek
comparing to h2vx output is a good start
#
tantek
if you want to black box it - which you don't have to - that's my point. h2vx is open source
#
gRegorLove
I really didn't want to dive into XSL
#
gRegorLove
Not sure what black box it means
gRegorLove_ joined the channel
#
tantek
black box it means to just look at something by what outputs does it give when you give it certain inputs, *without* looking at anything it does with those inputs to determine the outputs. it's a literal analogy. you put stuff into the black(opaque) box, you have no idea what happens inside, you get stuff out, then you reconstruct (theorize) what it did
#
tantek
prominent in reverse engineering
#
tantek
realizes he should have checked wikipedia
#
gRegorLove
So far the only differences is h2vx has language and charset for some content lines.
#
tantek
those are pretty essential
#
tantek
in particular, several calendars IIRC depend on explicit CHARSET to UTF-8 or else they use some rando charset (Latin-1?)
#
gRegorLove
Which I can add. I've got the 75 character wrapping with CRLF and the fields we need, which are minimal
#
tantek
originally just using my surname exposed a lot of that
#
tantek
but today, emojis in event names I'm sure would
#
gRegorLove
For the immediate use case (work project) I think we'll be cool as long as it supports iCal, gCal, and Outlook
#
gRegorLove
Yeah, I'm going to test with more utf8 characters
#
@dylanon
Now all set up to receive webmentions with Webmention.io. Will work on displaying them in context next. https://webmention.io
(twitter.com/_/status/889539734271340544)
snarfed, [miklb], KevinMarks, eli_oat1, j12t and KevinMarks_ joined the channel
#
@placerdiario
What the New Webmention and Annotation W3C Standards Mean for WordPress https://premium.wpmudev.org/blog/webmentions-annotations-web-standards/ WPMU DEV Blog
(twitter.com/_/status/889555280815685633)
snarfed joined the channel
#
voxpelli
Continuing from #indieweb: Looks like Lobster actually already fetches the full page: https://github.com/jcs/lobsters/blob/0ad6dffbe705f1d446461061f855a95b1df233d6/app/models/story.rb#L791 So would be easy to parse webmention endpoint from there
#
voxpelli
should maybe write a feature request after all
#
gRegorLove
What is Lobster?
#
Loqi
It looks like we don't have a page for "Lobster" yet. Would you like to create it?
#
gRegorLove
What is Lobsters?
#
Loqi
lobsters is a technology-focused link-aggregation site similar to Hacker News https://lobste.rs https://indieweb.org/lobsters
#
voxpelli
right, important -s there :P
#
gregorlove.com
edited /lobsters (+2) "link dfn"
(view diff)
#
gregorlove.com
edited /lobsters (+79) "open source, github link"
(view diff)
#
sebsel
aaronpk The XRay library release .zip has a .zip inside with the XRay library release
KevinMarks, KevinMarks_, tantek, j12t_, [chrisaldrich], snarfed and tbbrown joined the channel
#
aaronpk
sebsel: ?
#
aaronpk
i heard you like XRay so I put an XRay in ur XRay
#
tantek
hopes the .zip doens't have a .zip inside
#
tantek
that. shouldn't be possible
#
tantek
unless with each release it goes one deeper. inception style
#
aaronpk
Is that real
#
sebsel
I shopped one extra pane for the effect, because I already had that screenshot and threw the files out.
#
sebsel
But now I tried again, and yes, it's the same thing over and over again.
#
sknebel
seems like you included the previous release every time
#
sknebel
(now I want to figure out if you can make a zipquine like that, that actually is endless)
#
sebsel
Also, the views and the public folder are included too, so something is wrong there.
#
sknebel
(and .DS_STORE files, while we are complaining ;))
#
sebsel
Oh, yeah, it stopped after 5 zips :D
#
aaronpk
I have a script that zips the files in the current folder into the zip and I think it has been including the zip each time
#
aaronpk
I'll have to reevaluate my workflow :-)
#
sebsel
The previous xray-app.zip was 9.28 MB, the newest 15.5 MB, so maybe that one is affected too :P
#
KartikPrabhu
zipception
tantek joined the channel
#
Loqi
awesome
snarfed and KevinMarks joined the channel
#
@dylanon
Oops - Configured Webmention.io wrong. Tags must be added to all pages, not just the home page (as for IndieAuth).… https://dylanon.com/blog/2017/07/24/76357/
(twitter.com/_/status/889594647923838977)
KevinMarks_ and snarfed joined the channel
#
tantek.com
edited /Kirby (+178) "stub why section"
(view diff)
snarfed joined the channel
#
tantek.com
edited /Kirby (+0) "gen"
(view diff)
KevinMarks joined the channel
#
gregorlove.com
edited /Homebrew_Website_Club (+69) "/* Bellingham */ status"
(view diff)
#
tantek
snarfed: I'm trying to come up wtih a series of strategies for incrementally implementing showing responses of various types for various types of posts
#
tantek
to basically minimize unintended exposure to bad actors on my personal site
#
snarfed
yup. you mentioned starting with blacklisting all bridgy, to start. sgtm.
#
tantek
and in the process, figure out just what is required for at least some more "challenging" scenarios with showing backfeeds on your personal site
#
tantek
I realize most people here have (nearly) zero problem with their backfeeds, but I'd posit that's because of 1) small(er) amounts of attention, 2) the very strongly biased demographics of privilege (mostly white-ish and/or mostly male-ish)
#
tantek
because of @t incurring so much collateral backfeed damage, it's at least *a* more challenging use-case that may help explore (even address?) some of the problems beyond that privileged demographic.
#
tantek
trying to at least talk about it openly to explore it openly, and admit to the challenges along the way
#
tantek
responses from FB and IG are fairly high s/n, with rare abuse that not only is it quick to block, but the providers are quick to delete such accounts
#
tantek
oh that reminds me - does Bridgy do anything with backfeeds from accounts that services have taken down?
#
snarfed
it naively passes through the 404s etc
#
snarfed
nothing more explicit
#
tantek
there is the question of censorship enabling
#
tantek
however it may be too much to ask to try to discern when silos drop accounts for abuse vs other reasons
#
snarfed
more importantly, bridgy can't try to preserve speech that it hasn't previously fetched and stored
#
tantek
no that's not a problem
#
tantek
I mean with old speech from newly banned accounts
#
tantek
does that stay on Bridgy? or does Bridgy 410 those
#
snarfed
bridgy's mf2 conversion pretty much always fetches every time. it doesn't serve from storage
#
snarfed
so yes it would 404
#
tantek
oh wow had no idea those were fried not backed
#
tantek
baked*
#
aaronpk
fried not baked ?
#
tantek
aaronpk wait tell me you haven't heard that before
#
aaronpk
I've heard "baked" but not fried!
#
tantek
fried/baked analogy for backend page serving
#
tantek
wat seriously? this was the original contrast between WordPress and MT
#
aaronpk
well this is impossible to google
#
snarfed
re why, short answer is, we want to catch updates. https://github.com/snarfed/bridgy/issues/15#issuecomment-216609690
#
tantek
MT was baked (by perl into static files then served), WordPress always fried pages up as demanded from MySQL queries
#
Loqi
[snarfed] done, yay. oldest issue we've closed in a long time. :P oh and for the record, serving from stored responses would be tricky since we always want to serve the current silo response, or 404/410 if it's missing, etc. this should usually still work w...
#
tantek
makes sense snarfed
#
snarfed
on a related plumbing note, the current proposal for block lists in mf2 is ul.h-blocklist of li.h-cards. any objections? details: https://github.com/snarfed/granary/issues/40#issuecomment-266310635
#
Loqi
[snarfed] tentative proposed mf2: `ul.h-blocklist` of `li.h-card`s. based on http://microformats.org/wiki/block-list-brainstorming#simple_container_of_hCards (the `ul` could also have `h-feed`, but that may be redundant, and also the mf2 wiki says it's just ...
#
snarfed
or maybe we don't need any mf2, we just serve a list of h-cards, and application context determines that it's a block list
#
tantek
snarfed, in general we've avoided such "list" or "container" objects
#
tantek
because in practice they're not needed, unless the list or container has its own intrinsic properties (name, author, icon etc.)
#
Loqi
block-list-brainstorming
#
aaronpk
h-feed might be appropriate as a container for the h-cards tho?
#
tantek
so yeah, by default I would oppose an h-blocklist container without use-cases specifically for an explicit container object
#
aaronpk
oh yeah that page does seem to suggest a new top-level "blocklist" object
#
tantek
h-feed had to be justified by specific use-cases for it the list / container
#
aaronpk
"Proposed by Tantek during events/2008-08-12-weekly-meetup-dinner" omg almost 10 years ago
#
tantek
aaronpk of course it does because explicit is dev design by default
#
tantek
is vulnerable to that same flaw of course
#
tantek
being self aware of a flaw does not make you not vulnerable to it :)
#
tantek
ah the weekly uf dinners where we actually brainstormed and iterated on this stuff
#
aaronpk
weekly!
#
tantek
and local!
#
tantek
however nearly always only in SF
#
tantek
which is not really good for standards development frankly. far too insular / barriers for others
#
tantek
and you'll note (perhaps more meta) that I created wiki pages and upcoming events etc. for every one of them
#
tantek
has been feeling that pain for a while
#
tantek
(sometimes even Pownce and Plancast too when they were around)
#
eddiehinkle.com
edited /Homebrew_Website_Club (+179) "/* Cities getting started */ +Frederick, MD"
(view diff)
KevinMarks joined the channel
#
seblog.nl
edited /Kirby (+749) "/* Why */ expanding"
(view diff)
#
loqi.me
created /mass_adoption (+38) "prompted by tantek and dfn added by tantek"
(view diff)
#
sebsel
lol, then my post to glenn became on-topic for this discussion... I just got a spam-retweet-via-bridgy
#
gRegorLove
I might be seeing some attempts to webmention spam via the manual form at the bottom of my posts. Only on the php-mf2 anniversary post so far, though.
#
sebsel
what is the current thing to do? Manual block and delete?
#
Loqi
It looks like we don't have a page for "current thing to do" yet. Would you like to create it?
#
aaronpk
gRegorLove: i think that's just form spam
#
gRegorLove
The URLs don't even link to me, though. I added some more logging to confirm that's where they're coming from.
#
aaronpk
bots will submit any form they find
#
gRegorLove
Right, used the wrong term
#
snarfed
aaronpk++ yup. still no confirmed native wm spam in the wild yet afaik.
#
Loqi
aaronpk has 67 karma in this channel (1385 overall)
#
tantek
that being said, are there some basic form design (markup) techniques for catching at least the dumbest bots?
#
gRegorLove
I have a honeypot field which works pretty well, though I forgot to put it on that form.
#
tantek
like hidden input fields
#
tantek
what is a honeypot field
KevinMarks joined the channel
#
Loqi
It looks like we don't have a page for "honeypot field" yet. Would you like to create it?
#
gRegorLove
It's on my local comments form, borrowed from Known.
#
tantek
curious to see that technique documented with at least a link to an existing how to
#
aaronpk
i keep meaning to document that
#
aaronpk
i've used that for years, it works extremely well
#
gRegorLove
text field, display: none, placeholder text "If you are human, don't fill in this field."
#
loqi.me
created /Facebook_Comments_Plugin (+220) "prompted by tantek and dfn added by sknebel"
(view diff)
#
kaja.sknebel.net
edited /Facebook_Comments_Plugin (+1) "linkify ('… is <url>' pattern)"
(view diff)
#
tantek.com
edited /Facebook (+183) "/* Features */ User features Facebook handles on other sites:"
(view diff)
#
gRegorLove
honeypot field is a form input that is intended to be left blank and spam bots will usually fill in, allowing you to filter out a lot of spam form submissions.
#
loqi.me
created /honeypot_field (+186) "prompted by tantek and dfn added by gRegorLove"
(view diff)
#
tantek
what is a webmention form
#
Loqi
It looks like we don't have a page for "webmention form" yet. Would you like to create it?
#
gregorlove.com
edited /honeypot_field (+81) "IndieWeb Examples, See Also"
(view diff)
#
loqi.me
edited /Kirby (+235) "tantek added "2017-07-24 criticism: [http://glenn.thedixons.net/blog/giving-up-on-indieweb Giving Up on IndieWeb] <blockquote>Kirby plugins - I installed both micropub and webmentions - as far as I can tell, neither work as expected.</blockquote>" to"
(view diff)
#
gregorlove.com
edited /honeypot_field (+632) "/* IndieWeb Examples */ +me"
(view diff)
#
tantek.com
edited /Kirby (+324) "follow-up"
(view diff)
#
Loqi
[gRegor Morrill] Spam statistics since 2016-05-10, when I last reported: 1,224 caught by honeypot — an extra hidden field was filled in when it shouldn’t have been 828 marked as spam by Akismet 18 spam comments got through Akismet
#
tantek
Aside: folks who think certain UI things should "just be in the browser" ought to try writing one of these: http://webextensions-experiments.readthedocs.io/en/latest/
#
aaronpk
not in chrome yet?
#
aaronparecki.com
edited /p3k_naming_convention (+13) "/* Unused Names */"
(view diff)
#
aaronparecki.com
edited /p3k (-58) "/* Apps */"
(view diff)
tantek and snarfed joined the channel