#dev 2021-12-25

2021-12-25 UTC
jeremycherfas, sayanarijit[d], KartikPrabhu, tetov-irc and neceve joined the channel
#
capjamesg[d]
Is there a best practice for using a background image on a blog that does not take too long to load? I assume using a format like webp might be useful.
#
[tantek]
capjamesg[d], very small PNG, tiled
#
petermolnar
webp is overrated
#
sknebel
As always, performance is best answered by measuring. Check what the various png optimizers do, then see if if you get noticably better with something else
#
sknebel
(ok, almost always)
#
petermolnar
re longevity from meta: I've been thinking to somehow auto-switch to archive.org urls instead of normal links on my site after some years have passed. Trouble is, if I were to go out and check a url it could either be an error (so switch to archive.org) or a live link, but how do I verify if it's still the same content?
#
KartikPrabhu
one could save a small hash of the original content and compare? but that might fail for minor reasons like the CSS was updated or something
#
petermolnar
that can fail for many reasons
#
petermolnar
even if I were to save a quote with each url, the text could change if someone revisits it
#
petermolnar
but could still be the same article, except updated
#
[schmarty]
manualuntilithurts++
#
Loqi
manualuntilithurts has 1 karma over the last year
#
petermolnar
even the thought of visiting multiple thousands of urls hurts :)
#
[schmarty]
Even errors are not a perfect signal since a site can have temporary outages
#
[tantek]
“verify it's still the same content” <— 💁‍♂️🦋 is this IPFS?
#
[tantek]
KarthikPrabhu++ for identifying the most glaringly obvious fragility problem with the hashweb
#
Loqi
KarthikPrabhu has 1 karma over the last year
#
petermolnar
[tantek]: same-ish. Updated OK. Squat domain, no.
#
[tantek]
What is a squat domain
#
Loqi
It looks like we don't have a page for "squat domain" yet. Would you like to create it? (Or just say "squat domain is ____", a sentence describing the term)
#
[tantek]
What is a zombie
#
Loqi
zombie is in the context of the web a website that had died (site-deaths), perhaps due to domain registration neglect, and has been brought back by some other looking sorta like it did before, but oddly broken, often with spam pages/links added, and eats a lot of CPU likely due to abusive scripts https://indieweb.org/zombie
#
[tantek]
^ zombies are also a problem, where it vaguely looks like the thing you linked to but actually is now a spam farm
#
petermolnar
btw, a thought: I now know of quite a photographers, who had their domains and websites die off a long time ago, and the only place their work is still up is flickr.
#
petermolnar
I know one particular case where the reason he hadn't demolished his flickr is because he locked himself out: the email with the account is a former work email address of his. Thing is, I'm quite glad for this, given I love his work, and I'm quite a bit angry for stopping AND letting his sites rot.
#
petermolnar
At this point in time, there are many flickr accounts that lived significantly longer, than their counterpart websites, which makes me think that if a large amount of the netizens will eventually neglect their online content, silos might keep their work alive longer, than domains.
#
petermolnar
oh. I wanted this in -chat
#
petermolnar
although it's on topic
[davidmead] joined the channel
#
[davidmead]
!tell aaronpk I tried to post a webmention to https://aaronparecki.com/2018/06/30/11/your-first-webmention yesterday but it still hasn’t show up in the comments. My link seems good according to https://indiewebify.me & https://webmention.app. Not sure if I’m missing something
#
aaronpk
The one that says "webmention test"? I see it there
#
aaronpk
right under the "replies" header
KartikPrabhu joined the channel
#
[davidmead]
Great! Must’ve scrolled past it a couple of times now aaronpk. Saw the bottom entry was 2021 and thought mine would come after that 🙂
#
aaronpk
That page is getting really long
gRegor joined the channel
#
GWG
In a JSON Micropub submission, should visibility be an array with a single property?
#
Zegnat
goes to check spec
#
gRegor
I'm thinking the `p-name` should be removed from this how-to, similar to how on /note posts we no longer recommend p-name. Sound correct? https://indieweb.org/reply#Post_a_reply
#
gRegor
GWG: iirc, any property can be an array. IBC sends read-status as an array for example
#
GWG
gRegor: But in JSON notation, my code assumes it is always an array. So I just fixed that
#
Zegnat
GWG: if the property is not mp- prefixed, I think it should be an array, to be a valid mf2 JSON value. Otherwise I think your body might no longer be valid mf2 JSON
#
GWG
This came up with visibility
#
GWG
Which isn't mp prefixed
#
gRegor
I (temporarily) changed visibility to a string on IBC when someone had an issue, but I'm probably going to change it back: https://indiebookclub.biz/documentation
#
gRegor
(it ended up being a separate issue anyway, which I think was filed on WP Micropub plugin)
#
Zegnat
https://microformats.org/wiki/microformats2-json#properties is not an authorative spec (only the parser spec is) but should describe mf2 JSON correctly and says "Even when only one value is given, it will be inside an array."
#
GWG
gRegor: I just merged jeremyfelt's tests on the subject, and added a 'normalize' function that fixes it if incorrect
#
GWG
Incorrect as I define it.
#
GWG
So, it will now just work
#
Zegnat
IMO this is a weakness of the Micropub spec saying that it uses "the parsed Microformats 2 JSON format", but there is no formal definition of what that is. The closest is probably my mf2 json spec above and the matching JSON schema. But that is still waiting on an actual validation service before going anywhere.
#
GWG
Zegnat: Any movement on that?
#
Zegnat
I felt JSON schema validators for PHP were extremely bad at reporting errors thus making validation services all crap. At which point I gave up.
#
Zegnat
So I have the functioning schema done, but for now I am not planning to work on a service that handles validation
#
gRegor
Speaking of, I'm using micropublish.net to test my MP endpoint and looks like at least some requests are using strings instead of arrays.
#
gRegor
Looks like my code is handling it ok so far
#
GWG
Then my update was a good one.
#
gRegor
I like the preview feature there. Shows you the MP JSON
#
GWG
barryf++
#
Loqi
barryf has 1 karma in this channel over the last year (2 in all channels)
#
Zegnat
Do note that strings are fine in non-JSON pushes. So depending on if and how you keep those two apart, you might just find it easier to support both
#
aaronpk
mf2 json is defined in the microformats spec
#
GWG
Zegnat: I convert all form encoded submissions to json for handling anyway
#
aaronpk
the spec is literally how to build a JSON object from the microformats HTML https://microformats.org/wiki/microformats2-parsing#algorithm
#
Zegnat
I can just never find in the spec where it says which things should be arrays and stuff
#
Zegnat
Counter-intuitively I found it a lot easier to write a description of the final JSON, haha. But of course YMMV
#
GWG
Trying to decide what else I should tackle today. I did IndieAuth and Micropub tweaks.
#
gRegor
Time to add some unit tests to this module. yeah, yeah... should have started there XD
#
petermolnar
oh, that's glorious: href.li (wordpress.com & tumblr outgoing link body snatcher) doesn't do actual redirect, it's http-equiv="Refresh" and JS based.
#
gRegor
t.co does/did that too, bleh
#
Zegnat
I feel like thre was a UA trick for t.co to get by that. But yeah. The standard for those is to use HTML/JS redirects instead of HTTP redirects :(
#
Zegnat
Might even make sense. Not sure HTTP redirects would hide referer data
tetov-irc joined the channel