#dev 2020-05-09

2020-05-09 UTC
#
fuzzyBear
=)
#
fuzzyBear
out of curiosity, which one would be the better way? if that is a question, or which would be the better way of writting it
#
[tw2113]
i’d personally go with the second one with the chained functions
#
[tw2113]
i doubt one is really better than the other, so it’s more preference
#
fuzzyBear
next daft question(tbh I got loads, since Im self learning I find its difficult to get some clairty to some of these nuances that come accross the material
#
fuzzyBear
since JS is based on an Event Loop, which is a really complex for loop
#
fuzzyBear
and the engine is tuned for for loops, why does the language have other types of loops like while and do loops
#
fuzzyBear
one can throw your computer to searching for infinitiy and the other does nothing useful. so why have them
#
fuzzyBear
my next question is how do you know which method to use when you come across a problem, (this is problay more fundemental and would help me out, I find when I am looking at the problems and trying to figure out the solutions I just cant seem to find a way (yet) to get my solution into code
#
KartikPrabhu
we try to keep the discusison here specific to indieweb. Maybe #indieweb-chat channel is better for such general questions
[tantek] joined the channel
#
fuzzyBear
lol them lot told me to come here to ask questions
#
aaronpk
this isn't a general web development channel or community, we focus on personal websites
#
aaronpk
we're happy to help if this related to your personal website, but this isn't the best channel for general development questions
#
fuzzyBear
noted, sorry
#
KartikPrabhu
no need to apologise fuzzyBear, just keeping the discussion here in focus :)
xsteadfastx, [Katherine], [tw2113] and [chrisaldrich] joined the channel
#
jacky
okay so came back to it and made some changes to the target resolver
#
jacky
used your post as an example/test, aaronpk
#
jacky
as shown, only links in the posts are aimed to be targeted but everything else is still seen
#
jacky
sending webmentions to all of those by default feels like a bad idea to me
#
jacky
noticing that I might have a bug since it doesn't pick up the `in-reply-to` URL for tantek's post http://lighthouse.test.black.af/test?url=http%3A%2F%2Ftantek.com%2F2020%2F127%2Ft2%2F
#
jacky
oh nice, I def had a bug in my parser for this :)
#
dansup
hey jacky
#
jacky
test built and fixed
#
jacky
hey dansup; what's up?
#
jacky
I have! I was curious about it tbh but I have too many wheels spinning to stop :)
#
jacky
I can do it now tho
#
dansup
try following https://fedidb.org/actor/LRFKRSzEKg2lKYuoQCPN and check the logs ;)
#
dansup
trying to make AP easier for devs as well as test new features for Pixelfed
#
jacky
oh this like a managed solution for testing AP requests! nice
#
dansup
anyways, didn't mean to interrupt your wheels, have a good night bro!
#
jacky
nah lol breaks are critical :)
#
dansup
yeah its like a http request bin with a valid AP actor, I'm adding support for various AS verbs so you can test a wide variety of interactions
#
dansup
because it does federate, I need to keep track of instances to send federated deletes when the bin expires. Thats what I'm doing now
#
jacky
hmm I don't have an immediate way to do a follow
#
jacky
like outside of the mastodon UI
#
dansup
follows don't support mastodon yet, but it works with pleroma and pixelfed
#
dansup
mastodon has very strict validation lol
#
dansup
heh
#
dansup
If you have any advice or feedback for FediDB lmk!
#
jacky
I was going to ask how does one sign in but I guess that's invite-only?
#
dansup
nobody has joined yet, i'm still working on a few bugfixes before I open it up
#
jacky
outside of that, I'll keep it on my radar - caught me at a good time too, tomorrow's my "link read dump" day
#
jacky
I think aaronpk might be a ideal/good tester as he has a AP implementation/stand-in for his site
#
dansup
there are so many verbs, the validator has taken a lot of time 😅
#
jacky
I haven't worked on mine yet
#
jacky
is there a way to generate validation from the schema?
#
jacky
like a verb -> expectation parser of sorts?
#
dansup
you can post an object and validate or generate a [mastodon|pixelfed] verb
#
dansup
I'm hoping we can crowdsource more objects to generate
#
dansup
http sigs can be tricky too, I made a validator that debugs the most common problems
#
jacky
nice! I know that can be tricky esp since diff languages sort keys different
#
jacky
more tools like these are def needed in this space
#
jacky
I was talking to a friend and they mentioned what made it easy for them to be confident in using things like opengraph and twitter cards was the instant feedback testing tools they had
#
jacky
which is one of the high-effort, high-long-term-reward kind of things
nickodd, [tantek], strugee, [jgmac1106], gRegorLove, jeremych_, [jeremycherfas] and [Rose] joined the channel
#
Zegnat
Same is what I hear about Schema. The problem there being that the testing tool people always tell me about (Google’s) is not a generic Schema validator but instead a what-Google-Search-wants validator.
[KevinMarks] joined the channel
#
[KevinMarks]
well, that's the reason people are adding it in the first place.
#
Zegnat
Can’t dispute that, [KevinMarks], haha. Although I guess the same goes for OpenGraph and Twitter Cards, you can use the Facebook and Twitter validators respectively because you probably only care about those consumers.
#
[KevinMarks]
I agree that having a good validator loop matters; us having multiple ones can possibly confuse people. But as it is up to the consuming app how they present things I'm not sure how we avoid that. I'm the worst offender with unmung having a load of random presentations and no docs; indiewebify is the closest to a walkthrough
[LewisCowles] joined the channel
#
Loqi
Zegnat has 19 karma in this channel over the last year (55 in all channels)
#
[LewisCowles]
I won't dispute a majority as I have no data. But I personally used schema in the past because it supported a wide range of machine-parsable formats so that I could extract value from a page in-lieu of a database as part of front-to-back migrations, where the database is a known erroneous state, which is corrected before output, or contains vendor-specific hacks
#
[LewisCowles]
migrating an oracle database to anything is a test of patience at the data-layer because they like to crap as much incidental detail into their database solution as possible. A Frontend can supply a more utilitarian view
#
[LewisCowles]
and works okay for logical division / extracting user-domains. It's a bit misleading if you want domains to be owned by a business unit (BU)
#
[LewisCowles]
The above are reasons I disagree strongly with the "If you don't have a consumer" arguments
#
Zegnat
You used Schema as in schema.org properties in HTML? What machine-parsable formats does it support other than HTML? Or are you thinking of something else called Schema?
#
Zegnat
What is Schema?
#
Loqi
It looks like we don't have a page for "Schema" yet. Would you like to create it? (Or just say "Schema is ____", a sentence describing the term)
#
Zegnat
What is Schema.org?
#
Loqi
schema.org is a set of vocabularies created by a group at Google, in some cases based on existing vocabularies like microformats hCard, and in others re-inventing vocabularies instead of re-using specs like Atom or RSS, then co-launched in 2011 with fellow search-engine Microsoft Bing, and their search licensee Yahoo, still run by them and additional search engine Yandex https://indieweb.org/schema.org
#
[LewisCowles]
I used schema.org in-case it were needed for frontend to backend migration. Every system I've had to front-to-back migrate was manually processed using template inspection, which is shorter and cheaper than diving into oracle, but more expensive than needing to parse out from a schema that is pre-defined. It's a middle ground saving a fair bit of time with a carrot google might like them, but that being a tangential benefit rather than the
#
[LewisCowles]
If you crap your reviews into a soup of divs and spans, it's fairly likely it might be very problematic or even manual to transfer them at a later date to a system. If you use schema.org you know you're covered for most SaaS providers properties, which can easily be transformed or read by those providers. Fairly certain that Trustpilot took >10,000 reviews from a website I worked on by parsing the frontend
#
[LewisCowles]
It's not an argument for schema.org specifically, but rather any non-not-invented-here frontend markup with semi-logical structure
#
[LewisCowles]
a larger problem is in systems with non-uniform unit sizes
#
[LewisCowles]
5 may be 5 stars or 5/10
#
Zegnat
That is why microformats (and I think schema.org too?) has properties for clarifying what your range is, IIRC
#
[LewisCowles]
sure and had I known about microformats I may have used
#
[LewisCowles]
this was over 6 years ago
#
[LewisCowles]
mf likely existed, but wasn't on my radar and even now seems to have less market share
#
[LewisCowles]
market share can also be thought-share. If something has a million users, it could be likely to support more edges than something with 5 users (not that either figure reflects these two)
[jgmac1106] joined the channel
#
[LewisCowles]
On my personal website I'll use things I'd never deploy to a customer. That is a pity, but both are machine-readable inline. One goes from data-format -> text -> data-format, the other although text is already in data-format representation. I can see why people use it. I understand and to some degree buy-into not repeating yourself.
markopasha joined the channel
#
[jgmac1106]
making img-maps in photoshop actually quite straight forward, do wish I knew more about the css only pop up modals I have been seeing be nice to combine
[grantcodes] joined the channel
#
[grantcodes]
Probably a dialog element? They can work without js but they're better with js
#
[LewisCowles]
for more information on dialogs https://github.com/whatwg/html/issues/3567 may help. then again it may be awful
#
Loqi
[keithamus] #3567 Have some way of opening `<dialog>` elements without JavaScript
#
[LewisCowles]
don't be misled by the unfurl, lots of in-depth discussion of JS and non-JS methods
#
[jgmac1106]
ha ha jeremycherfas the photoshop img splicing doesn't use an img-map at all...it is old school you get a table with each link being a gif in a specific cell
#
[KevinMarks]
Ah, you don't want that.
#
[jgmac1106]
no, not all
#
[jgmac1106]
I was trying photoshop to avoid that..
#
[KevinMarks]
Make the image map by hand, you just need coords
#
[KevinMarks]
Or draw hot areas over your image in an svg editor and link them.
#
[jgmac1106]
yeah I can still use photoshop to get the gridlines, was thinking of trying svg, need to learn more vector graphics anyways
#
Loqi
svg has 1 karma in this channel over the last year (3 in all channels)
#
Loqi
kevinmarks has 13 karma in this channel over the last year (54 in all channels)
Henster1 and markopasha joined the channel
#
jeremycherfas
Oh well; live and learn.
[snarfed] and [jeremycherfas] joined the channel
#
[LewisCowles]
does anyone else have a clusertHAT for raspberry Pi?
#
[LewisCowles]
Mine keeps turning off p1 when p4 is booted and p4 when p1 is booted from USB. Fairly annoying
nickodd and [jgarber] joined the channel
#
[LewisCowles]
I think it's power related. I'll plug the cluster HAT into a power only USB
#
[LewisCowles]
well. I'll try that anyway. It's on a 2.4A 5V power supply
markopasha and flex14 joined the channel
#
lahacker
following up from last week's wc hwc chat regarding browser extensions i've followed the rabbit hole full circle (?) to a dead reference at https://wiki.mozilla.org/HCard to the `Operator` of yesterdecade https://web.archive.org/web/20110610050725/https://addons.mozilla.org/en-US/firefox/addon/operator/
#
lahacker
does anyone remember using it? would that be a desirable thing to have today?
#
lahacker
reading up on browser extension "best practices" https://extensionworkshop.com/documentation/develop/user-experience-best-practices/ i see the "page action" is the most appropriate UX for the case of finding microformats on the current page
#
Loqi
mdnwebdocs-bot
xsteadfastx joined the channel
#
Zegnat
yes page action would be the one to use.
#
Zegnat
I remember using Operator in the mf1 days, but I never really had a use for it. Was more of a fun thing for me back then.
#
lahacker
zegnat yeah same but it was a different time
#
lahacker
the most recent discussions revolve around leveraging in-page <indie-action>s which seems kind of kludgy from a UX perspective
#
lahacker
like i have to look around for the place in the page that had the original reply button just to find an incompatibly style button
#
lahacker
maybe there's a use case i'm not seeing
#
Zegnat
I think indie-action was specifically for letting the website have buttons in their own styling that would trigger an action in your browser/on your site.
#
Zegnat
If you want to interact with a site from your browser rather than from the site UI, there are tools for that already
#
Zegnat
What is Omnibear?
#
lahacker
yeah you're right -- web hero just went about that one wrong
#
Loqi
Omnibear is a browser extension for posting text notes, replies, and likes to your website using Micropub https://indieweb.org/Omnibear
#
Zegnat
That is one such example. Omnibear I believe parses the page you are on to grant you quick access to doing something with it on your site (through micropub).
#
Zegnat
Omnibear uses microformats parsing for that, I believe. And does it irregardless of the website’s own UI, it lives in your browser, like Operator before.
#
lahacker
ahh yes this is the missing piece
#
Zegnat
Take that one for a spin lahacker :) I do not off the top of my head know any other mf2 parsing indieweb related browser extensions
nickodd left the channel
#
Zegnat
There is one for rel=me checking that [KevinMarks] and I worked on, but that has a very limited scope
markopasha joined the channel
#
lahacker
zegnat++
#
Loqi
zegnat has 19 karma in this channel over the last year (57 in all channels)
#
lahacker
good stuff in there -- going to come up with some hybrid of them all!
[KevinMarks] joined the channel
#
[KevinMarks]
There's the fragmention one too, birthdays also very limited
markopasha, KartikPrabhu, [juju], hs0ucy, xsteadfastx and [chrisaldrich] joined the channel
#
Zegnat
Anyone have any ideas what the difference might be in Travis running phpunit on PHP 7.3, and me running it locally? I have a couple of tests that I am passing locally but that Travis fails: https://github.com/microformats/php-mf2/pull/163#issuecomment-626231088 compare https://travis-ci.org/github/microformats/php-mf2/jobs/685139785
swentel, [tantek], KartikPrabhu, [KevinMarks] and sudomake joined the channel
#
jmac
Webmention storage question: If a friend sends a wm to http://jmac.org, and (after a forward) https://jmac.org is happy to receive it, what is the recorded target of that wm?
#
Loqi
Jason McIntosh
#
jmac
Yes, thank you loqi
#
jmac
Anyway, in the past I've cheated by surreptitiously editing the stored http://... into https://..., but that doesn't seem ideal
sudomake joined the channel
#
jmac
Also not ideal is to understand a query for "Show me all mentions for https://jmac.org" as "Show me all mentions for https://jmac.org and also every URL that might forward to it"
#
Loqi
Jason McIntosh
#
jmac
thanks loqi, sharp as a tack as always
#
Loqi
you're welcome
#
swentel
I don't store the full domain, but just the path
#
swentel
so in this case I'd store '/'
#
sknebel
I consider webmentions to be attached to the final target, but do store the target that was sent so I can re-verify easily
vika_nezrimaya joined the channel
#
sknebel
I think it makes sense for a query to cover all redirected urls also, but the original data should be available
#
sudomake
hello, I am getting a white page after upgrading to php 7.3.
[schmarty] joined the channel
#
jmac
sknebel: Yes, just storing both the literal, sender-specified target URL and the ultimate-actual-receiver URL sounds reasonable
#
jmac
And that would have nice parity with storing both the literal source and the "original source" as is made necessary by e.g. Bridgy webmentions
gRegorLove, markopasha, sudomake and KartikPrabhu joined the channel
#
jmac
Now I have a refreshingly basic question about h-cards
#
jmac
I wish to add an h-card to (almost) every page of jmac.org, and I figured the best way to do that would be to just sneak it into the sidebar on the site's template, which already has my name printed on it anyway. So far, so good
#
jmac
However, according to https://indiewebify.me/validate-h-card/?url=https%3A%2F%2Fjmac.org%2Ftech-writing-portfolio%2F (which links to one arbitrary page on my website that's not the front page), this isn't quite right, and the h-card on every page is supposed to have a u-url property matching the page it's on -- it's not enough to just refer to my website's front page. Is that correct?
#
jmac
I certainly can do that easily enough using my templating system, but it makes me wonder if my whole concept about what h-cards are even for isn't quite right somehow
#
sknebel
jmac: note that it says "check your *homepage* h-card"
#
sknebel
that should be a representative h-card
#
sknebel
that test doesn't apply for individual entries page
sudomake joined the channel
#
jmac
So it's not an error to have identical h-card HTML (every one referring to a u-url of "/") on every sub-page of my site?
#
sknebel
it's not really necessary, but not an error
#
jmac
Agree it's not really necessary, but I'm messing around with posting note-replies to friends, and in this way discovered that webmentions sourced from pages within my site had no author information attached, how embarrassing
sudomake and flex14 joined the channel
#
[tantek]
[Katherine]’s request yesterday in #indieweb got me thinking that there's likely some UI analysis that's worth doing, on SMS /create UIs, and share UIs in general
#
[tantek]
in particular these parts of [Katherine]’s request: "create blog posts via SMS", "a website for me and my friends to share links to articles under a common domain", "i want my friends to be able to send texts with a URL and a short description to accompany the post then it'll appear on the site like a light weight blog [post]"
#
[tantek]
I think "create blog posts via SMS" is a reasonable request and also speaks to the need for simpler posting UIs that don't confront you with all the fields etc., or auth-flow etc.
sudomake left the channel
#
[KevinMarks]
Quill has micropub by email, so potentially could use sms to email for that
#
[schmarty]
Posting by SMS really takes me back!
#
[tantek]
What's the attraction in posting by SMS?
#
[schmarty]
That's gonna vary person to person for sure. Back when I co-built an MMS-to-photoblog social sharing site it was the only "output" available on most mobile devices!
#
[tantek]
simple mobile posting
#
[schmarty]
These days I'd guess the appeal is it's a distraction-free, extremely minimal UI for making low-effort posts
#
[schmarty]
OH! Also SMS meant you had an "identifier" for the author. Nowadays we recognize how insecure that is but it was powerful not to have to log in to things
#
[schmarty]
On boarding new people to the photoblog was just a matter of giving them the email address to MMS at. Their source address (phone number, basically) would identify the sender.
[snarfed] joined the channel
#
Loqi
signups has -1 karma over the last year
#
[schmarty]
signups--